Last active
May 15, 2016 05:34
-
-
Save stas-sl/a952a550d7b24eece29647e8c8567da9 to your computer and use it in GitHub Desktop.
This file has been truncated, but you can view the full file.
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| I0510 15:24:57.106338 10926 solver.cpp:280] Solving mixed_lstm | |
| I0510 15:24:57.106350 10926 solver.cpp:281] Learning Rate Policy: fixed | |
| I0510 15:24:57.414443 10926 solver.cpp:229] Iteration 0, loss = 28.4505 | |
| I0510 15:24:57.414511 10926 solver.cpp:245] Train net output #0: loss1/accuracy = 0 | |
| I0510 15:24:57.414530 10926 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0510 15:24:57.414542 10926 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0510 15:24:57.414554 10926 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0510 15:24:57.414566 10926 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0 | |
| I0510 15:24:57.414578 10926 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0 | |
| I0510 15:24:57.414590 10926 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0 | |
| I0510 15:24:57.414602 10926 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0 | |
| I0510 15:24:57.414614 10926 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0 | |
| I0510 15:24:57.414626 10926 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0 | |
| I0510 15:24:57.414638 10926 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0 | |
| I0510 15:24:57.414650 10926 solver.cpp:245] Train net output #11: loss1/accuracy11 = 0 | |
| I0510 15:24:57.414662 10926 solver.cpp:245] Train net output #12: loss1/accuracy12 = 0 | |
| I0510 15:24:57.414674 10926 solver.cpp:245] Train net output #13: loss1/accuracy13 = 0 | |
| I0510 15:24:57.414686 10926 solver.cpp:245] Train net output #14: loss1/accuracy14 = 0.125 | |
| I0510 15:24:57.414698 10926 solver.cpp:245] Train net output #15: loss1/accuracy15 = 0 | |
| I0510 15:24:57.414710 10926 solver.cpp:245] Train net output #16: loss1/accuracy16 = 0.25 | |
| I0510 15:24:57.414722 10926 solver.cpp:245] Train net output #17: loss1/accuracy17 = 0 | |
| I0510 15:24:57.414734 10926 solver.cpp:245] Train net output #18: loss1/accuracy18 = 0 | |
| I0510 15:24:57.414746 10926 solver.cpp:245] Train net output #19: loss1/accuracy19 = 0 | |
| I0510 15:24:57.414758 10926 solver.cpp:245] Train net output #20: loss1/accuracy20 = 0 | |
| I0510 15:24:57.414769 10926 solver.cpp:245] Train net output #21: loss1/accuracy21 = 0 | |
| I0510 15:24:57.414809 10926 solver.cpp:245] Train net output #22: loss1/accuracy22 = 0 | |
| I0510 15:24:57.414824 10926 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0 | |
| I0510 15:24:57.414835 10926 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0 | |
| I0510 15:24:57.414855 10926 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 4.36235 (* 0.3 = 1.3087 loss) | |
| I0510 15:24:57.414871 10926 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 4.31106 (* 0.3 = 1.29332 loss) | |
| I0510 15:24:57.414886 10926 solver.cpp:245] Train net output #27: loss1/loss01 = 5.46726 (* 0.0272727 = 0.149107 loss) | |
| I0510 15:24:57.414901 10926 solver.cpp:245] Train net output #28: loss1/loss02 = 4.73896 (* 0.0272727 = 0.129244 loss) | |
| I0510 15:24:57.414914 10926 solver.cpp:245] Train net output #29: loss1/loss03 = 4.65407 (* 0.0272727 = 0.126929 loss) | |
| I0510 15:24:57.414929 10926 solver.cpp:245] Train net output #30: loss1/loss04 = 4.20216 (* 0.0272727 = 0.114604 loss) | |
| I0510 15:24:57.414943 10926 solver.cpp:245] Train net output #31: loss1/loss05 = 4.5028 (* 0.0272727 = 0.122804 loss) | |
| I0510 15:24:57.414959 10926 solver.cpp:245] Train net output #32: loss1/loss06 = 4.28253 (* 0.0272727 = 0.116796 loss) | |
| I0510 15:24:57.414975 10926 solver.cpp:245] Train net output #33: loss1/loss07 = 4.64458 (* 0.0272727 = 0.12667 loss) | |
| I0510 15:24:57.414990 10926 solver.cpp:245] Train net output #34: loss1/loss08 = 6.32047 (* 0.0272727 = 0.172377 loss) | |
| I0510 15:24:57.415005 10926 solver.cpp:245] Train net output #35: loss1/loss09 = 5.19835 (* 0.0272727 = 0.141773 loss) | |
| I0510 15:24:57.415019 10926 solver.cpp:245] Train net output #36: loss1/loss10 = 4.87704 (* 0.0272727 = 0.13301 loss) | |
| I0510 15:24:57.415033 10926 solver.cpp:245] Train net output #37: loss1/loss11 = 4.61487 (* 0.0272727 = 0.12586 loss) | |
| I0510 15:24:57.415048 10926 solver.cpp:245] Train net output #38: loss1/loss12 = 4.38139 (* 0.0272727 = 0.119492 loss) | |
| I0510 15:24:57.415062 10926 solver.cpp:245] Train net output #39: loss1/loss13 = 4.89324 (* 0.0272727 = 0.133452 loss) | |
| I0510 15:24:57.415076 10926 solver.cpp:245] Train net output #40: loss1/loss14 = 3.68531 (* 0.0272727 = 0.100508 loss) | |
| I0510 15:24:57.415091 10926 solver.cpp:245] Train net output #41: loss1/loss15 = 4.62144 (* 0.0272727 = 0.126039 loss) | |
| I0510 15:24:57.415104 10926 solver.cpp:245] Train net output #42: loss1/loss16 = 3.48895 (* 0.0272727 = 0.0951532 loss) | |
| I0510 15:24:57.415118 10926 solver.cpp:245] Train net output #43: loss1/loss17 = 4.43237 (* 0.0272727 = 0.120883 loss) | |
| I0510 15:24:57.415133 10926 solver.cpp:245] Train net output #44: loss1/loss18 = 5.12372 (* 0.0272727 = 0.139738 loss) | |
| I0510 15:24:57.415146 10926 solver.cpp:245] Train net output #45: loss1/loss19 = 4.64013 (* 0.0272727 = 0.126549 loss) | |
| I0510 15:24:57.415161 10926 solver.cpp:245] Train net output #46: loss1/loss20 = 4.74878 (* 0.0272727 = 0.129512 loss) | |
| I0510 15:24:57.415175 10926 solver.cpp:245] Train net output #47: loss1/loss21 = 4.61697 (* 0.0272727 = 0.125917 loss) | |
| I0510 15:24:57.415189 10926 solver.cpp:245] Train net output #48: loss1/loss22 = 4.37096 (* 0.0272727 = 0.119208 loss) | |
| I0510 15:24:57.415201 10926 solver.cpp:245] Train net output #49: loss2/accuracy = 0 | |
| I0510 15:24:57.415213 10926 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0510 15:24:57.415225 10926 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0510 15:24:57.415237 10926 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0510 15:24:57.415248 10926 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0 | |
| I0510 15:24:57.415261 10926 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0 | |
| I0510 15:24:57.415272 10926 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0 | |
| I0510 15:24:57.415283 10926 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0 | |
| I0510 15:24:57.415295 10926 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0 | |
| I0510 15:24:57.415307 10926 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0 | |
| I0510 15:24:57.415328 10926 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0 | |
| I0510 15:24:57.415343 10926 solver.cpp:245] Train net output #60: loss2/accuracy11 = 0 | |
| I0510 15:24:57.415354 10926 solver.cpp:245] Train net output #61: loss2/accuracy12 = 0 | |
| I0510 15:24:57.415366 10926 solver.cpp:245] Train net output #62: loss2/accuracy13 = 0 | |
| I0510 15:24:57.415377 10926 solver.cpp:245] Train net output #63: loss2/accuracy14 = 0.125 | |
| I0510 15:24:57.415393 10926 solver.cpp:245] Train net output #64: loss2/accuracy15 = 0 | |
| I0510 15:24:57.415405 10926 solver.cpp:245] Train net output #65: loss2/accuracy16 = 0 | |
| I0510 15:24:57.415417 10926 solver.cpp:245] Train net output #66: loss2/accuracy17 = 0 | |
| I0510 15:24:57.415429 10926 solver.cpp:245] Train net output #67: loss2/accuracy18 = 0 | |
| I0510 15:24:57.415441 10926 solver.cpp:245] Train net output #68: loss2/accuracy19 = 0 | |
| I0510 15:24:57.415452 10926 solver.cpp:245] Train net output #69: loss2/accuracy20 = 0 | |
| I0510 15:24:57.415463 10926 solver.cpp:245] Train net output #70: loss2/accuracy21 = 0 | |
| I0510 15:24:57.415475 10926 solver.cpp:245] Train net output #71: loss2/accuracy22 = 0 | |
| I0510 15:24:57.415487 10926 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0 | |
| I0510 15:24:57.415498 10926 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.0243902 | |
| I0510 15:24:57.415513 10926 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 4.3466 (* 0.3 = 1.30398 loss) | |
| I0510 15:24:57.415524 10926 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 4.28574 (* 0.3 = 1.28572 loss) | |
| I0510 15:24:57.415534 10926 solver.cpp:245] Train net output #76: loss2/loss01 = 4.46762 (* 0.0272727 = 0.121844 loss) | |
| I0510 15:24:57.415547 10926 solver.cpp:245] Train net output #77: loss2/loss02 = 4.68519 (* 0.0272727 = 0.127778 loss) | |
| I0510 15:24:57.415561 10926 solver.cpp:245] Train net output #78: loss2/loss03 = 4.22211 (* 0.0272727 = 0.115148 loss) | |
| I0510 15:24:57.415576 10926 solver.cpp:245] Train net output #79: loss2/loss04 = 4.14911 (* 0.0272727 = 0.113158 loss) | |
| I0510 15:24:57.415591 10926 solver.cpp:245] Train net output #80: loss2/loss05 = 4.38952 (* 0.0272727 = 0.119714 loss) | |
| I0510 15:24:57.415604 10926 solver.cpp:245] Train net output #81: loss2/loss06 = 4.62692 (* 0.0272727 = 0.126189 loss) | |
| I0510 15:24:57.415618 10926 solver.cpp:245] Train net output #82: loss2/loss07 = 4.50004 (* 0.0272727 = 0.122728 loss) | |
| I0510 15:24:57.415633 10926 solver.cpp:245] Train net output #83: loss2/loss08 = 4.39117 (* 0.0272727 = 0.119759 loss) | |
| I0510 15:24:57.415647 10926 solver.cpp:245] Train net output #84: loss2/loss09 = 4.5434 (* 0.0272727 = 0.123911 loss) | |
| I0510 15:24:57.415662 10926 solver.cpp:245] Train net output #85: loss2/loss10 = 4.39337 (* 0.0272727 = 0.119819 loss) | |
| I0510 15:24:57.415675 10926 solver.cpp:245] Train net output #86: loss2/loss11 = 5.26425 (* 0.0272727 = 0.14357 loss) | |
| I0510 15:24:57.415689 10926 solver.cpp:245] Train net output #87: loss2/loss12 = 4.55749 (* 0.0272727 = 0.124295 loss) | |
| I0510 15:24:57.415704 10926 solver.cpp:245] Train net output #88: loss2/loss13 = 4.5836 (* 0.0272727 = 0.125007 loss) | |
| I0510 15:24:57.415717 10926 solver.cpp:245] Train net output #89: loss2/loss14 = 4.39572 (* 0.0272727 = 0.119883 loss) | |
| I0510 15:24:57.415732 10926 solver.cpp:245] Train net output #90: loss2/loss15 = 4.39666 (* 0.0272727 = 0.119909 loss) | |
| I0510 15:24:57.415746 10926 solver.cpp:245] Train net output #91: loss2/loss16 = 4.21805 (* 0.0272727 = 0.115038 loss) | |
| I0510 15:24:57.415760 10926 solver.cpp:245] Train net output #92: loss2/loss17 = 4.87184 (* 0.0272727 = 0.132868 loss) | |
| I0510 15:24:57.415774 10926 solver.cpp:245] Train net output #93: loss2/loss18 = 4.62084 (* 0.0272727 = 0.126023 loss) | |
| I0510 15:24:57.415788 10926 solver.cpp:245] Train net output #94: loss2/loss19 = 5.03362 (* 0.0272727 = 0.137281 loss) | |
| I0510 15:24:57.415802 10926 solver.cpp:245] Train net output #95: loss2/loss20 = 4.76918 (* 0.0272727 = 0.130069 loss) | |
| I0510 15:24:57.415827 10926 solver.cpp:245] Train net output #96: loss2/loss21 = 4.66158 (* 0.0272727 = 0.127134 loss) | |
| I0510 15:24:57.415843 10926 solver.cpp:245] Train net output #97: loss2/loss22 = 4.75508 (* 0.0272727 = 0.129684 loss) | |
| I0510 15:24:57.415854 10926 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0243902 | |
| I0510 15:24:57.415868 10926 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0510 15:24:57.415879 10926 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.125 | |
| I0510 15:24:57.415891 10926 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0510 15:24:57.415902 10926 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0 | |
| I0510 15:24:57.415915 10926 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0 | |
| I0510 15:24:57.415926 10926 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0 | |
| I0510 15:24:57.415937 10926 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0 | |
| I0510 15:24:57.415949 10926 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0 | |
| I0510 15:24:57.415961 10926 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0 | |
| I0510 15:24:57.415972 10926 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0 | |
| I0510 15:24:57.415984 10926 solver.cpp:245] Train net output #109: loss3/accuracy11 = 0 | |
| I0510 15:24:57.415995 10926 solver.cpp:245] Train net output #110: loss3/accuracy12 = 0 | |
| I0510 15:24:57.416009 10926 solver.cpp:245] Train net output #111: loss3/accuracy13 = 0 | |
| I0510 15:24:57.416021 10926 solver.cpp:245] Train net output #112: loss3/accuracy14 = 0.125 | |
| I0510 15:24:57.416033 10926 solver.cpp:245] Train net output #113: loss3/accuracy15 = 0 | |
| I0510 15:24:57.416045 10926 solver.cpp:245] Train net output #114: loss3/accuracy16 = 0 | |
| I0510 15:24:57.416056 10926 solver.cpp:245] Train net output #115: loss3/accuracy17 = 0 | |
| I0510 15:24:57.416069 10926 solver.cpp:245] Train net output #116: loss3/accuracy18 = 0 | |
| I0510 15:24:57.416079 10926 solver.cpp:245] Train net output #117: loss3/accuracy19 = 0 | |
| I0510 15:24:57.416090 10926 solver.cpp:245] Train net output #118: loss3/accuracy20 = 0 | |
| I0510 15:24:57.416101 10926 solver.cpp:245] Train net output #119: loss3/accuracy21 = 0 | |
| I0510 15:24:57.416113 10926 solver.cpp:245] Train net output #120: loss3/accuracy22 = 0 | |
| I0510 15:24:57.416126 10926 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.00568182 | |
| I0510 15:24:57.416137 10926 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.0243902 | |
| I0510 15:24:57.416152 10926 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 4.27167 (* 1 = 4.27167 loss) | |
| I0510 15:24:57.416165 10926 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 4.5745 (* 1 = 4.5745 loss) | |
| I0510 15:24:57.416179 10926 solver.cpp:245] Train net output #125: loss3/loss01 = 4.39413 (* 0.0909091 = 0.399467 loss) | |
| I0510 15:24:57.416193 10926 solver.cpp:245] Train net output #126: loss3/loss02 = 3.99239 (* 0.0909091 = 0.362945 loss) | |
| I0510 15:24:57.416210 10926 solver.cpp:245] Train net output #127: loss3/loss03 = 4.22973 (* 0.0909091 = 0.384521 loss) | |
| I0510 15:24:57.416221 10926 solver.cpp:245] Train net output #128: loss3/loss04 = 4.46985 (* 0.0909091 = 0.40635 loss) | |
| I0510 15:24:57.416235 10926 solver.cpp:245] Train net output #129: loss3/loss05 = 4.27994 (* 0.0909091 = 0.389085 loss) | |
| I0510 15:24:57.416250 10926 solver.cpp:245] Train net output #130: loss3/loss06 = 4.32544 (* 0.0909091 = 0.393222 loss) | |
| I0510 15:24:57.416265 10926 solver.cpp:245] Train net output #131: loss3/loss07 = 4.56383 (* 0.0909091 = 0.414893 loss) | |
| I0510 15:24:57.416278 10926 solver.cpp:245] Train net output #132: loss3/loss08 = 4.82644 (* 0.0909091 = 0.438767 loss) | |
| I0510 15:24:57.416292 10926 solver.cpp:245] Train net output #133: loss3/loss09 = 4.8338 (* 0.0909091 = 0.439436 loss) | |
| I0510 15:24:57.416307 10926 solver.cpp:245] Train net output #134: loss3/loss10 = 3.61173 (* 0.0909091 = 0.328339 loss) | |
| I0510 15:24:57.416332 10926 solver.cpp:245] Train net output #135: loss3/loss11 = 4.64854 (* 0.0909091 = 0.422594 loss) | |
| I0510 15:24:57.416347 10926 solver.cpp:245] Train net output #136: loss3/loss12 = 4.58389 (* 0.0909091 = 0.416717 loss) | |
| I0510 15:24:57.416363 10926 solver.cpp:245] Train net output #137: loss3/loss13 = 4.40369 (* 0.0909091 = 0.400335 loss) | |
| I0510 15:24:57.416376 10926 solver.cpp:245] Train net output #138: loss3/loss14 = 3.89031 (* 0.0909091 = 0.353665 loss) | |
| I0510 15:24:57.416390 10926 solver.cpp:245] Train net output #139: loss3/loss15 = 4.2804 (* 0.0909091 = 0.389127 loss) | |
| I0510 15:24:57.416405 10926 solver.cpp:245] Train net output #140: loss3/loss16 = 4.92609 (* 0.0909091 = 0.447826 loss) | |
| I0510 15:24:57.416419 10926 solver.cpp:245] Train net output #141: loss3/loss17 = 4.40941 (* 0.0909091 = 0.400855 loss) | |
| I0510 15:24:57.416436 10926 solver.cpp:245] Train net output #142: loss3/loss18 = 4.58048 (* 0.0909091 = 0.416408 loss) | |
| I0510 15:24:57.416451 10926 solver.cpp:245] Train net output #143: loss3/loss19 = 4.339 (* 0.0909091 = 0.394454 loss) | |
| I0510 15:24:57.416466 10926 solver.cpp:245] Train net output #144: loss3/loss20 = 5.43763 (* 0.0909091 = 0.49433 loss) | |
| I0510 15:24:57.416481 10926 solver.cpp:245] Train net output #145: loss3/loss21 = 4.01123 (* 0.0909091 = 0.364658 loss) | |
| I0510 15:24:57.416494 10926 solver.cpp:245] Train net output #146: loss3/loss22 = 4.59974 (* 0.0909091 = 0.418158 loss) | |
| I0510 15:24:57.416507 10926 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0510 15:24:57.416518 10926 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0510 15:24:57.416529 10926 solver.cpp:245] Train net output #149: total_confidence = 1.37478e-35 | |
| I0510 15:24:57.416543 10926 solver.cpp:245] Train net output #150: total_confidence_not_rec = 7.26934e-32 | |
| I0510 15:24:57.416563 10926 sgd_solver.cpp:106] Iteration 0, lr = 0.001 | |
| I0510 15:27:24.889616 10926 solver.cpp:229] Iteration 500, loss = 14.7737 | |
| I0510 15:27:24.890060 10926 solver.cpp:245] Train net output #0: loss1/accuracy = 0 | |
| I0510 15:27:24.890082 10926 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0510 15:27:24.890095 10926 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0510 15:27:24.890107 10926 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0510 15:27:24.890120 10926 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.25 | |
| I0510 15:27:24.890132 10926 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.125 | |
| I0510 15:27:24.890146 10926 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.25 | |
| I0510 15:27:24.890157 10926 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.75 | |
| I0510 15:27:24.890169 10926 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.875 | |
| I0510 15:27:24.890182 10926 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.875 | |
| I0510 15:27:24.890194 10926 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.875 | |
| I0510 15:27:24.890215 10926 solver.cpp:245] Train net output #11: loss1/accuracy11 = 0.875 | |
| I0510 15:27:24.890228 10926 solver.cpp:245] Train net output #12: loss1/accuracy12 = 0.875 | |
| I0510 15:27:24.890239 10926 solver.cpp:245] Train net output #13: loss1/accuracy13 = 0.875 | |
| I0510 15:27:24.890254 10926 solver.cpp:245] Train net output #14: loss1/accuracy14 = 0.875 | |
| I0510 15:27:24.890265 10926 solver.cpp:245] Train net output #15: loss1/accuracy15 = 0.875 | |
| I0510 15:27:24.890277 10926 solver.cpp:245] Train net output #16: loss1/accuracy16 = 0.875 | |
| I0510 15:27:24.890296 10926 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0510 15:27:24.890308 10926 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0510 15:27:24.890321 10926 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0510 15:27:24.890336 10926 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0510 15:27:24.890357 10926 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0510 15:27:24.890372 10926 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0510 15:27:24.890390 10926 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.6875 | |
| I0510 15:27:24.890403 10926 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.145455 | |
| I0510 15:27:24.890419 10926 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 4.28729 (* 0.3 = 1.28619 loss) | |
| I0510 15:27:24.890434 10926 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.91244 (* 0.3 = 0.573732 loss) | |
| I0510 15:27:24.890449 10926 solver.cpp:245] Train net output #27: loss1/loss01 = 3.8225 (* 0.0272727 = 0.10425 loss) | |
| I0510 15:27:24.890468 10926 solver.cpp:245] Train net output #28: loss1/loss02 = 4.42431 (* 0.0272727 = 0.120663 loss) | |
| I0510 15:27:24.890482 10926 solver.cpp:245] Train net output #29: loss1/loss03 = 4.24342 (* 0.0272727 = 0.11573 loss) | |
| I0510 15:27:24.890496 10926 solver.cpp:245] Train net output #30: loss1/loss04 = 3.60244 (* 0.0272727 = 0.0982483 loss) | |
| I0510 15:27:24.890511 10926 solver.cpp:245] Train net output #31: loss1/loss05 = 3.92116 (* 0.0272727 = 0.106941 loss) | |
| I0510 15:27:24.890525 10926 solver.cpp:245] Train net output #32: loss1/loss06 = 3.94805 (* 0.0272727 = 0.107674 loss) | |
| I0510 15:27:24.890539 10926 solver.cpp:245] Train net output #33: loss1/loss07 = 1.85285 (* 0.0272727 = 0.0505322 loss) | |
| I0510 15:27:24.890554 10926 solver.cpp:245] Train net output #34: loss1/loss08 = 1.09676 (* 0.0272727 = 0.0299117 loss) | |
| I0510 15:27:24.890568 10926 solver.cpp:245] Train net output #35: loss1/loss09 = 0.815202 (* 0.0272727 = 0.0222328 loss) | |
| I0510 15:27:24.890583 10926 solver.cpp:245] Train net output #36: loss1/loss10 = 1.23058 (* 0.0272727 = 0.0335614 loss) | |
| I0510 15:27:24.890596 10926 solver.cpp:245] Train net output #37: loss1/loss11 = 0.968782 (* 0.0272727 = 0.0264213 loss) | |
| I0510 15:27:24.890610 10926 solver.cpp:245] Train net output #38: loss1/loss12 = 1.22067 (* 0.0272727 = 0.0332911 loss) | |
| I0510 15:27:24.890625 10926 solver.cpp:245] Train net output #39: loss1/loss13 = 1.00897 (* 0.0272727 = 0.0275173 loss) | |
| I0510 15:27:24.890657 10926 solver.cpp:245] Train net output #40: loss1/loss14 = 0.783499 (* 0.0272727 = 0.0213682 loss) | |
| I0510 15:27:24.890673 10926 solver.cpp:245] Train net output #41: loss1/loss15 = 0.757411 (* 0.0272727 = 0.0206567 loss) | |
| I0510 15:27:24.890687 10926 solver.cpp:245] Train net output #42: loss1/loss16 = 1.04768 (* 0.0272727 = 0.0285731 loss) | |
| I0510 15:27:24.890702 10926 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0872958 (* 0.0272727 = 0.0023808 loss) | |
| I0510 15:27:24.890717 10926 solver.cpp:245] Train net output #44: loss1/loss18 = 0.14028 (* 0.0272727 = 0.00382581 loss) | |
| I0510 15:27:24.890730 10926 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0593406 (* 0.0272727 = 0.00161838 loss) | |
| I0510 15:27:24.890745 10926 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0471335 (* 0.0272727 = 0.00128546 loss) | |
| I0510 15:27:24.890760 10926 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0595725 (* 0.0272727 = 0.0016247 loss) | |
| I0510 15:27:24.890774 10926 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0627516 (* 0.0272727 = 0.00171141 loss) | |
| I0510 15:27:24.890787 10926 solver.cpp:245] Train net output #49: loss2/accuracy = 0 | |
| I0510 15:27:24.890799 10926 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.125 | |
| I0510 15:27:24.890812 10926 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0510 15:27:24.890821 10926 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0510 15:27:24.890835 10926 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0 | |
| I0510 15:27:24.890854 10926 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.125 | |
| I0510 15:27:24.890869 10926 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.25 | |
| I0510 15:27:24.890894 10926 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.75 | |
| I0510 15:27:24.890907 10926 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.875 | |
| I0510 15:27:24.890918 10926 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.875 | |
| I0510 15:27:24.890930 10926 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.875 | |
| I0510 15:27:24.890943 10926 solver.cpp:245] Train net output #60: loss2/accuracy11 = 0.875 | |
| I0510 15:27:24.890954 10926 solver.cpp:245] Train net output #61: loss2/accuracy12 = 0.875 | |
| I0510 15:27:24.890975 10926 solver.cpp:245] Train net output #62: loss2/accuracy13 = 0.875 | |
| I0510 15:27:24.890987 10926 solver.cpp:245] Train net output #63: loss2/accuracy14 = 0.875 | |
| I0510 15:27:24.891000 10926 solver.cpp:245] Train net output #64: loss2/accuracy15 = 0.875 | |
| I0510 15:27:24.891012 10926 solver.cpp:245] Train net output #65: loss2/accuracy16 = 0.875 | |
| I0510 15:27:24.891024 10926 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0510 15:27:24.891036 10926 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0510 15:27:24.891048 10926 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0510 15:27:24.891059 10926 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0510 15:27:24.891075 10926 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0510 15:27:24.891088 10926 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0510 15:27:24.891099 10926 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.6875 | |
| I0510 15:27:24.891113 10926 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.127273 | |
| I0510 15:27:24.891126 10926 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 4.29527 (* 0.3 = 1.28858 loss) | |
| I0510 15:27:24.891140 10926 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.88665 (* 0.3 = 0.565996 loss) | |
| I0510 15:27:24.891155 10926 solver.cpp:245] Train net output #76: loss2/loss01 = 4.47345 (* 0.0272727 = 0.122003 loss) | |
| I0510 15:27:24.891170 10926 solver.cpp:245] Train net output #77: loss2/loss02 = 4.21013 (* 0.0272727 = 0.114822 loss) | |
| I0510 15:27:24.891196 10926 solver.cpp:245] Train net output #78: loss2/loss03 = 4.85562 (* 0.0272727 = 0.132426 loss) | |
| I0510 15:27:24.891216 10926 solver.cpp:245] Train net output #79: loss2/loss04 = 4.33767 (* 0.0272727 = 0.1183 loss) | |
| I0510 15:27:24.891232 10926 solver.cpp:245] Train net output #80: loss2/loss05 = 4.4398 (* 0.0272727 = 0.121086 loss) | |
| I0510 15:27:24.891245 10926 solver.cpp:245] Train net output #81: loss2/loss06 = 4.31632 (* 0.0272727 = 0.117718 loss) | |
| I0510 15:27:24.891259 10926 solver.cpp:245] Train net output #82: loss2/loss07 = 1.6242 (* 0.0272727 = 0.0442964 loss) | |
| I0510 15:27:24.891273 10926 solver.cpp:245] Train net output #83: loss2/loss08 = 1.02427 (* 0.0272727 = 0.0279346 loss) | |
| I0510 15:27:24.891295 10926 solver.cpp:245] Train net output #84: loss2/loss09 = 0.940278 (* 0.0272727 = 0.025644 loss) | |
| I0510 15:27:24.891309 10926 solver.cpp:245] Train net output #85: loss2/loss10 = 0.773854 (* 0.0272727 = 0.0211051 loss) | |
| I0510 15:27:24.891324 10926 solver.cpp:245] Train net output #86: loss2/loss11 = 1.09303 (* 0.0272727 = 0.0298099 loss) | |
| I0510 15:27:24.891338 10926 solver.cpp:245] Train net output #87: loss2/loss12 = 0.873733 (* 0.0272727 = 0.0238291 loss) | |
| I0510 15:27:24.891352 10926 solver.cpp:245] Train net output #88: loss2/loss13 = 0.714014 (* 0.0272727 = 0.0194731 loss) | |
| I0510 15:27:24.891366 10926 solver.cpp:245] Train net output #89: loss2/loss14 = 1.05187 (* 0.0272727 = 0.0286875 loss) | |
| I0510 15:27:24.891381 10926 solver.cpp:245] Train net output #90: loss2/loss15 = 0.901481 (* 0.0272727 = 0.0245858 loss) | |
| I0510 15:27:24.891394 10926 solver.cpp:245] Train net output #91: loss2/loss16 = 1.05186 (* 0.0272727 = 0.0286872 loss) | |
| I0510 15:27:24.891408 10926 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0678201 (* 0.0272727 = 0.00184964 loss) | |
| I0510 15:27:24.891423 10926 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0603541 (* 0.0272727 = 0.00164602 loss) | |
| I0510 15:27:24.891438 10926 solver.cpp:245] Train net output #94: loss2/loss19 = 0.0611336 (* 0.0272727 = 0.00166728 loss) | |
| I0510 15:27:24.891453 10926 solver.cpp:245] Train net output #95: loss2/loss20 = 0.0713926 (* 0.0272727 = 0.00194707 loss) | |
| I0510 15:27:24.891466 10926 solver.cpp:245] Train net output #96: loss2/loss21 = 0.128577 (* 0.0272727 = 0.00350664 loss) | |
| I0510 15:27:24.891480 10926 solver.cpp:245] Train net output #97: loss2/loss22 = 0.054703 (* 0.0272727 = 0.0014919 loss) | |
| I0510 15:27:24.891494 10926 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0363636 | |
| I0510 15:27:24.891505 10926 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0510 15:27:24.891517 10926 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0510 15:27:24.891530 10926 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0510 15:27:24.891543 10926 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0510 15:27:24.891566 10926 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.125 | |
| I0510 15:27:24.891585 10926 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.25 | |
| I0510 15:27:24.891599 10926 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.75 | |
| I0510 15:27:24.891610 10926 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.875 | |
| I0510 15:27:24.891623 10926 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.875 | |
| I0510 15:27:24.891634 10926 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.875 | |
| I0510 15:27:24.891646 10926 solver.cpp:245] Train net output #109: loss3/accuracy11 = 0.875 | |
| I0510 15:27:24.891659 10926 solver.cpp:245] Train net output #110: loss3/accuracy12 = 0.875 | |
| I0510 15:27:24.891669 10926 solver.cpp:245] Train net output #111: loss3/accuracy13 = 0.875 | |
| I0510 15:27:24.891681 10926 solver.cpp:245] Train net output #112: loss3/accuracy14 = 0.875 | |
| I0510 15:27:24.891693 10926 solver.cpp:245] Train net output #113: loss3/accuracy15 = 0.875 | |
| I0510 15:27:24.891705 10926 solver.cpp:245] Train net output #114: loss3/accuracy16 = 0.875 | |
| I0510 15:27:24.891728 10926 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0510 15:27:24.891741 10926 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0510 15:27:24.891753 10926 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0510 15:27:24.891765 10926 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0510 15:27:24.891777 10926 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0510 15:27:24.891788 10926 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0510 15:27:24.891799 10926 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.6875 | |
| I0510 15:27:24.891811 10926 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.218182 | |
| I0510 15:27:24.891826 10926 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 4.0035 (* 1 = 4.0035 loss) | |
| I0510 15:27:24.891839 10926 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.47202 (* 1 = 1.47202 loss) | |
| I0510 15:27:24.891854 10926 solver.cpp:245] Train net output #125: loss3/loss01 = 4.31965 (* 0.0909091 = 0.392695 loss) | |
| I0510 15:27:24.891868 10926 solver.cpp:245] Train net output #126: loss3/loss02 = 4.11154 (* 0.0909091 = 0.373776 loss) | |
| I0510 15:27:24.891881 10926 solver.cpp:245] Train net output #127: loss3/loss03 = 4.03258 (* 0.0909091 = 0.366598 loss) | |
| I0510 15:27:24.891896 10926 solver.cpp:245] Train net output #128: loss3/loss04 = 3.8697 (* 0.0909091 = 0.351791 loss) | |
| I0510 15:27:24.891909 10926 solver.cpp:245] Train net output #129: loss3/loss05 = 3.75066 (* 0.0909091 = 0.340969 loss) | |
| I0510 15:27:24.891923 10926 solver.cpp:245] Train net output #130: loss3/loss06 = 3.27308 (* 0.0909091 = 0.297552 loss) | |
| I0510 15:27:24.891940 10926 solver.cpp:245] Train net output #131: loss3/loss07 = 1.81715 (* 0.0909091 = 0.165196 loss) | |
| I0510 15:27:24.891954 10926 solver.cpp:245] Train net output #132: loss3/loss08 = 1.0043 (* 0.0909091 = 0.0912997 loss) | |
| I0510 15:27:24.891968 10926 solver.cpp:245] Train net output #133: loss3/loss09 = 1.00136 (* 0.0909091 = 0.091033 loss) | |
| I0510 15:27:24.891983 10926 solver.cpp:245] Train net output #134: loss3/loss10 = 0.785341 (* 0.0909091 = 0.0713947 loss) | |
| I0510 15:27:24.891996 10926 solver.cpp:245] Train net output #135: loss3/loss11 = 0.721258 (* 0.0909091 = 0.0655689 loss) | |
| I0510 15:27:24.892010 10926 solver.cpp:245] Train net output #136: loss3/loss12 = 1.16467 (* 0.0909091 = 0.105879 loss) | |
| I0510 15:27:24.892024 10926 solver.cpp:245] Train net output #137: loss3/loss13 = 0.908524 (* 0.0909091 = 0.0825931 loss) | |
| I0510 15:27:24.892037 10926 solver.cpp:245] Train net output #138: loss3/loss14 = 1.09803 (* 0.0909091 = 0.0998207 loss) | |
| I0510 15:27:24.892051 10926 solver.cpp:245] Train net output #139: loss3/loss15 = 0.877906 (* 0.0909091 = 0.0798097 loss) | |
| I0510 15:27:24.892066 10926 solver.cpp:245] Train net output #140: loss3/loss16 = 1.1071 (* 0.0909091 = 0.100645 loss) | |
| I0510 15:27:24.892079 10926 solver.cpp:245] Train net output #141: loss3/loss17 = 0.0245946 (* 0.0909091 = 0.00223588 loss) | |
| I0510 15:27:24.892093 10926 solver.cpp:245] Train net output #142: loss3/loss18 = 0.0184597 (* 0.0909091 = 0.00167815 loss) | |
| I0510 15:27:24.892107 10926 solver.cpp:245] Train net output #143: loss3/loss19 = 0.0156215 (* 0.0909091 = 0.00142014 loss) | |
| I0510 15:27:24.892130 10926 solver.cpp:245] Train net output #144: loss3/loss20 = 0.0104825 (* 0.0909091 = 0.000952951 loss) | |
| I0510 15:27:24.892158 10926 solver.cpp:245] Train net output #145: loss3/loss21 = 0.0146713 (* 0.0909091 = 0.00133376 loss) | |
| I0510 15:27:24.892174 10926 solver.cpp:245] Train net output #146: loss3/loss22 = 0.0139693 (* 0.0909091 = 0.00126994 loss) | |
| I0510 15:27:24.892185 10926 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0510 15:27:24.892197 10926 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0510 15:27:24.892213 10926 solver.cpp:245] Train net output #149: total_confidence = 1.01629e-08 | |
| I0510 15:27:24.892236 10926 solver.cpp:245] Train net output #150: total_confidence_not_rec = 5.86186e-07 | |
| I0510 15:27:24.892252 10926 sgd_solver.cpp:106] Iteration 500, lr = 0.001 | |
| I0510 15:29:52.711612 10926 solver.cpp:229] Iteration 1000, loss = 13.409 | |
| I0510 15:29:52.711760 10926 solver.cpp:245] Train net output #0: loss1/accuracy = 0 | |
| I0510 15:29:52.711781 10926 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.125 | |
| I0510 15:29:52.711794 10926 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.125 | |
| I0510 15:29:52.711807 10926 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0510 15:29:52.711818 10926 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.125 | |
| I0510 15:29:52.711832 10926 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.25 | |
| I0510 15:29:52.711843 10926 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.5 | |
| I0510 15:29:52.711855 10926 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.625 | |
| I0510 15:29:52.711869 10926 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.625 | |
| I0510 15:29:52.711884 10926 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.625 | |
| I0510 15:29:52.711896 10926 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.75 | |
| I0510 15:29:52.711910 10926 solver.cpp:245] Train net output #11: loss1/accuracy11 = 0.875 | |
| I0510 15:29:52.711922 10926 solver.cpp:245] Train net output #12: loss1/accuracy12 = 0.875 | |
| I0510 15:29:52.711935 10926 solver.cpp:245] Train net output #13: loss1/accuracy13 = 0.875 | |
| I0510 15:29:52.711946 10926 solver.cpp:245] Train net output #14: loss1/accuracy14 = 0.875 | |
| I0510 15:29:52.711958 10926 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0510 15:29:52.711971 10926 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0510 15:29:52.711983 10926 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0510 15:29:52.711995 10926 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0510 15:29:52.712007 10926 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0510 15:29:52.712018 10926 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0510 15:29:52.712030 10926 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0510 15:29:52.712043 10926 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0510 15:29:52.712054 10926 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.676136 | |
| I0510 15:29:52.712066 10926 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.0877193 | |
| I0510 15:29:52.712082 10926 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 4.3075 (* 0.3 = 1.29225 loss) | |
| I0510 15:29:52.712098 10926 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.75385 (* 0.3 = 0.526155 loss) | |
| I0510 15:29:52.712112 10926 solver.cpp:245] Train net output #27: loss1/loss01 = 4.04854 (* 0.0272727 = 0.110415 loss) | |
| I0510 15:29:52.712126 10926 solver.cpp:245] Train net output #28: loss1/loss02 = 4.41979 (* 0.0272727 = 0.12054 loss) | |
| I0510 15:29:52.712141 10926 solver.cpp:245] Train net output #29: loss1/loss03 = 3.71893 (* 0.0272727 = 0.101425 loss) | |
| I0510 15:29:52.712155 10926 solver.cpp:245] Train net output #30: loss1/loss04 = 4.26287 (* 0.0272727 = 0.11626 loss) | |
| I0510 15:29:52.712169 10926 solver.cpp:245] Train net output #31: loss1/loss05 = 3.53744 (* 0.0272727 = 0.0964757 loss) | |
| I0510 15:29:52.712183 10926 solver.cpp:245] Train net output #32: loss1/loss06 = 3.16925 (* 0.0272727 = 0.086434 loss) | |
| I0510 15:29:52.712198 10926 solver.cpp:245] Train net output #33: loss1/loss07 = 2.13839 (* 0.0272727 = 0.0583197 loss) | |
| I0510 15:29:52.712211 10926 solver.cpp:245] Train net output #34: loss1/loss08 = 1.99956 (* 0.0272727 = 0.0545335 loss) | |
| I0510 15:29:52.712225 10926 solver.cpp:245] Train net output #35: loss1/loss09 = 2.40354 (* 0.0272727 = 0.0655511 loss) | |
| I0510 15:29:52.712239 10926 solver.cpp:245] Train net output #36: loss1/loss10 = 1.55155 (* 0.0272727 = 0.042315 loss) | |
| I0510 15:29:52.712254 10926 solver.cpp:245] Train net output #37: loss1/loss11 = 1.08925 (* 0.0272727 = 0.0297069 loss) | |
| I0510 15:29:52.712267 10926 solver.cpp:245] Train net output #38: loss1/loss12 = 0.996508 (* 0.0272727 = 0.0271775 loss) | |
| I0510 15:29:52.712281 10926 solver.cpp:245] Train net output #39: loss1/loss13 = 1.11693 (* 0.0272727 = 0.0304616 loss) | |
| I0510 15:29:52.712313 10926 solver.cpp:245] Train net output #40: loss1/loss14 = 0.900904 (* 0.0272727 = 0.0245701 loss) | |
| I0510 15:29:52.712330 10926 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0834805 (* 0.0272727 = 0.00227674 loss) | |
| I0510 15:29:52.712344 10926 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0309244 (* 0.0272727 = 0.000843393 loss) | |
| I0510 15:29:52.712359 10926 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0257262 (* 0.0272727 = 0.000701622 loss) | |
| I0510 15:29:52.712373 10926 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0347817 (* 0.0272727 = 0.000948593 loss) | |
| I0510 15:29:52.712388 10926 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0382091 (* 0.0272727 = 0.00104207 loss) | |
| I0510 15:29:52.712402 10926 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0251925 (* 0.0272727 = 0.000687067 loss) | |
| I0510 15:29:52.712416 10926 solver.cpp:245] Train net output #47: loss1/loss21 = 0.036635 (* 0.0272727 = 0.000999136 loss) | |
| I0510 15:29:52.712430 10926 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0222326 (* 0.0272727 = 0.000606343 loss) | |
| I0510 15:29:52.712442 10926 solver.cpp:245] Train net output #49: loss2/accuracy = 0 | |
| I0510 15:29:52.712455 10926 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0510 15:29:52.712466 10926 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0510 15:29:52.712478 10926 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0510 15:29:52.712491 10926 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.125 | |
| I0510 15:29:52.712502 10926 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.125 | |
| I0510 15:29:52.712514 10926 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.5 | |
| I0510 15:29:52.712527 10926 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.625 | |
| I0510 15:29:52.712538 10926 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.625 | |
| I0510 15:29:52.712550 10926 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.625 | |
| I0510 15:29:52.712563 10926 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.75 | |
| I0510 15:29:52.712574 10926 solver.cpp:245] Train net output #60: loss2/accuracy11 = 0.875 | |
| I0510 15:29:52.712586 10926 solver.cpp:245] Train net output #61: loss2/accuracy12 = 0.875 | |
| I0510 15:29:52.712599 10926 solver.cpp:245] Train net output #62: loss2/accuracy13 = 0.875 | |
| I0510 15:29:52.712610 10926 solver.cpp:245] Train net output #63: loss2/accuracy14 = 0.875 | |
| I0510 15:29:52.712622 10926 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0510 15:29:52.712635 10926 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0510 15:29:52.712646 10926 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0510 15:29:52.712657 10926 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0510 15:29:52.712669 10926 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0510 15:29:52.712682 10926 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0510 15:29:52.712692 10926 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0510 15:29:52.712704 10926 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0510 15:29:52.712716 10926 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.676136 | |
| I0510 15:29:52.712728 10926 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.122807 | |
| I0510 15:29:52.712743 10926 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 4.16278 (* 0.3 = 1.24883 loss) | |
| I0510 15:29:52.712756 10926 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.77949 (* 0.3 = 0.533846 loss) | |
| I0510 15:29:52.712770 10926 solver.cpp:245] Train net output #76: loss2/loss01 = 3.59627 (* 0.0272727 = 0.09808 loss) | |
| I0510 15:29:52.712787 10926 solver.cpp:245] Train net output #77: loss2/loss02 = 4.10821 (* 0.0272727 = 0.112042 loss) | |
| I0510 15:29:52.712815 10926 solver.cpp:245] Train net output #78: loss2/loss03 = 3.83814 (* 0.0272727 = 0.104676 loss) | |
| I0510 15:29:52.712829 10926 solver.cpp:245] Train net output #79: loss2/loss04 = 4.14622 (* 0.0272727 = 0.113079 loss) | |
| I0510 15:29:52.712843 10926 solver.cpp:245] Train net output #80: loss2/loss05 = 4.35739 (* 0.0272727 = 0.118838 loss) | |
| I0510 15:29:52.712857 10926 solver.cpp:245] Train net output #81: loss2/loss06 = 3.21444 (* 0.0272727 = 0.0876667 loss) | |
| I0510 15:29:52.712872 10926 solver.cpp:245] Train net output #82: loss2/loss07 = 1.97096 (* 0.0272727 = 0.0537535 loss) | |
| I0510 15:29:52.712885 10926 solver.cpp:245] Train net output #83: loss2/loss08 = 2.21094 (* 0.0272727 = 0.0602983 loss) | |
| I0510 15:29:52.712899 10926 solver.cpp:245] Train net output #84: loss2/loss09 = 2.16789 (* 0.0272727 = 0.0591243 loss) | |
| I0510 15:29:52.712913 10926 solver.cpp:245] Train net output #85: loss2/loss10 = 2.01058 (* 0.0272727 = 0.054834 loss) | |
| I0510 15:29:52.712929 10926 solver.cpp:245] Train net output #86: loss2/loss11 = 1.01853 (* 0.0272727 = 0.0277781 loss) | |
| I0510 15:29:52.712944 10926 solver.cpp:245] Train net output #87: loss2/loss12 = 0.763334 (* 0.0272727 = 0.0208182 loss) | |
| I0510 15:29:52.712959 10926 solver.cpp:245] Train net output #88: loss2/loss13 = 0.990939 (* 0.0272727 = 0.0270256 loss) | |
| I0510 15:29:52.712972 10926 solver.cpp:245] Train net output #89: loss2/loss14 = 0.885337 (* 0.0272727 = 0.0241456 loss) | |
| I0510 15:29:52.712987 10926 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0519862 (* 0.0272727 = 0.0014178 loss) | |
| I0510 15:29:52.713001 10926 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0452432 (* 0.0272727 = 0.00123391 loss) | |
| I0510 15:29:52.713016 10926 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0404571 (* 0.0272727 = 0.00110338 loss) | |
| I0510 15:29:52.713030 10926 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0259405 (* 0.0272727 = 0.000707469 loss) | |
| I0510 15:29:52.713044 10926 solver.cpp:245] Train net output #94: loss2/loss19 = 0.0207971 (* 0.0272727 = 0.000567193 loss) | |
| I0510 15:29:52.713058 10926 solver.cpp:245] Train net output #95: loss2/loss20 = 0.0340019 (* 0.0272727 = 0.000927324 loss) | |
| I0510 15:29:52.713073 10926 solver.cpp:245] Train net output #96: loss2/loss21 = 0.0267166 (* 0.0272727 = 0.000728635 loss) | |
| I0510 15:29:52.713086 10926 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0249853 (* 0.0272727 = 0.000681416 loss) | |
| I0510 15:29:52.713099 10926 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0175439 | |
| I0510 15:29:52.713110 10926 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0510 15:29:52.713140 10926 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.125 | |
| I0510 15:29:52.713153 10926 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0.25 | |
| I0510 15:29:52.713165 10926 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0 | |
| I0510 15:29:52.713177 10926 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.25 | |
| I0510 15:29:52.713189 10926 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.5 | |
| I0510 15:29:52.713201 10926 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.625 | |
| I0510 15:29:52.713213 10926 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.625 | |
| I0510 15:29:52.713224 10926 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.625 | |
| I0510 15:29:52.713237 10926 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.75 | |
| I0510 15:29:52.713248 10926 solver.cpp:245] Train net output #109: loss3/accuracy11 = 0.875 | |
| I0510 15:29:52.713259 10926 solver.cpp:245] Train net output #110: loss3/accuracy12 = 0.875 | |
| I0510 15:29:52.713271 10926 solver.cpp:245] Train net output #111: loss3/accuracy13 = 0.875 | |
| I0510 15:29:52.713284 10926 solver.cpp:245] Train net output #112: loss3/accuracy14 = 0.875 | |
| I0510 15:29:52.713295 10926 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0510 15:29:52.713306 10926 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0510 15:29:52.713331 10926 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0510 15:29:52.713345 10926 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0510 15:29:52.713356 10926 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0510 15:29:52.713367 10926 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0510 15:29:52.713379 10926 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0510 15:29:52.713390 10926 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0510 15:29:52.713402 10926 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.670455 | |
| I0510 15:29:52.713414 10926 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.0877193 | |
| I0510 15:29:52.713428 10926 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.99148 (* 1 = 3.99148 loss) | |
| I0510 15:29:52.713443 10926 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.58431 (* 1 = 1.58431 loss) | |
| I0510 15:29:52.713456 10926 solver.cpp:245] Train net output #125: loss3/loss01 = 3.82121 (* 0.0909091 = 0.347383 loss) | |
| I0510 15:29:52.713470 10926 solver.cpp:245] Train net output #126: loss3/loss02 = 3.57306 (* 0.0909091 = 0.324823 loss) | |
| I0510 15:29:52.713484 10926 solver.cpp:245] Train net output #127: loss3/loss03 = 3.58933 (* 0.0909091 = 0.326302 loss) | |
| I0510 15:29:52.713497 10926 solver.cpp:245] Train net output #128: loss3/loss04 = 3.8387 (* 0.0909091 = 0.348973 loss) | |
| I0510 15:29:52.713511 10926 solver.cpp:245] Train net output #129: loss3/loss05 = 3.48447 (* 0.0909091 = 0.31677 loss) | |
| I0510 15:29:52.713526 10926 solver.cpp:245] Train net output #130: loss3/loss06 = 2.69099 (* 0.0909091 = 0.244635 loss) | |
| I0510 15:29:52.713538 10926 solver.cpp:245] Train net output #131: loss3/loss07 = 2.09427 (* 0.0909091 = 0.190388 loss) | |
| I0510 15:29:52.713552 10926 solver.cpp:245] Train net output #132: loss3/loss08 = 1.9979 (* 0.0909091 = 0.181627 loss) | |
| I0510 15:29:52.713567 10926 solver.cpp:245] Train net output #133: loss3/loss09 = 2.10813 (* 0.0909091 = 0.191648 loss) | |
| I0510 15:29:52.713580 10926 solver.cpp:245] Train net output #134: loss3/loss10 = 1.78344 (* 0.0909091 = 0.162131 loss) | |
| I0510 15:29:52.713594 10926 solver.cpp:245] Train net output #135: loss3/loss11 = 0.910519 (* 0.0909091 = 0.0827745 loss) | |
| I0510 15:29:52.713608 10926 solver.cpp:245] Train net output #136: loss3/loss12 = 0.748322 (* 0.0909091 = 0.0680293 loss) | |
| I0510 15:29:52.713623 10926 solver.cpp:245] Train net output #137: loss3/loss13 = 0.814567 (* 0.0909091 = 0.0740515 loss) | |
| I0510 15:29:52.713636 10926 solver.cpp:245] Train net output #138: loss3/loss14 = 0.827204 (* 0.0909091 = 0.0752004 loss) | |
| I0510 15:29:52.713650 10926 solver.cpp:245] Train net output #139: loss3/loss15 = 0.0380271 (* 0.0909091 = 0.00345701 loss) | |
| I0510 15:29:52.713665 10926 solver.cpp:245] Train net output #140: loss3/loss16 = 0.0338639 (* 0.0909091 = 0.00307854 loss) | |
| I0510 15:29:52.713678 10926 solver.cpp:245] Train net output #141: loss3/loss17 = 0.0139578 (* 0.0909091 = 0.00126889 loss) | |
| I0510 15:29:52.713692 10926 solver.cpp:245] Train net output #142: loss3/loss18 = 0.0128029 (* 0.0909091 = 0.0011639 loss) | |
| I0510 15:29:52.713706 10926 solver.cpp:245] Train net output #143: loss3/loss19 = 0.0108394 (* 0.0909091 = 0.000985399 loss) | |
| I0510 15:29:52.713721 10926 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00949374 (* 0.0909091 = 0.000863068 loss) | |
| I0510 15:29:52.713734 10926 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00607524 (* 0.0909091 = 0.000552294 loss) | |
| I0510 15:29:52.713747 10926 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00893123 (* 0.0909091 = 0.00081193 loss) | |
| I0510 15:29:52.713760 10926 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0510 15:29:52.713771 10926 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0510 15:29:52.713783 10926 solver.cpp:245] Train net output #149: total_confidence = 2.68905e-10 | |
| I0510 15:29:52.713804 10926 solver.cpp:245] Train net output #150: total_confidence_not_rec = 5.59668e-07 | |
| I0510 15:29:52.713819 10926 sgd_solver.cpp:106] Iteration 1000, lr = 0.001 | |
| I0510 15:32:19.685997 10926 solver.cpp:229] Iteration 1500, loss = 13.101 | |
| I0510 15:32:19.686151 10926 solver.cpp:245] Train net output #0: loss1/accuracy = 0 | |
| I0510 15:32:19.686170 10926 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0510 15:32:19.686178 10926 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0510 15:32:19.686187 10926 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.125 | |
| I0510 15:32:19.686194 10926 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0 | |
| I0510 15:32:19.686203 10926 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.25 | |
| I0510 15:32:19.686210 10926 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.5 | |
| I0510 15:32:19.686218 10926 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.75 | |
| I0510 15:32:19.686225 10926 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.875 | |
| I0510 15:32:19.686233 10926 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0510 15:32:19.686240 10926 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0510 15:32:19.686249 10926 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0510 15:32:19.686256 10926 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0510 15:32:19.686264 10926 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0510 15:32:19.686271 10926 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0510 15:32:19.686278 10926 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0510 15:32:19.686285 10926 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0510 15:32:19.686293 10926 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0510 15:32:19.686300 10926 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0510 15:32:19.686307 10926 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0510 15:32:19.686314 10926 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0510 15:32:19.686322 10926 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0510 15:32:19.686331 10926 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0510 15:32:19.686337 10926 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.75 | |
| I0510 15:32:19.686345 10926 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.0909091 | |
| I0510 15:32:19.686357 10926 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 4.4423 (* 0.3 = 1.33269 loss) | |
| I0510 15:32:19.686367 10926 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.50827 (* 0.3 = 0.452481 loss) | |
| I0510 15:32:19.686378 10926 solver.cpp:245] Train net output #27: loss1/loss01 = 4.22197 (* 0.0272727 = 0.115145 loss) | |
| I0510 15:32:19.686388 10926 solver.cpp:245] Train net output #28: loss1/loss02 = 4.38229 (* 0.0272727 = 0.119517 loss) | |
| I0510 15:32:19.686396 10926 solver.cpp:245] Train net output #29: loss1/loss03 = 4.65871 (* 0.0272727 = 0.127056 loss) | |
| I0510 15:32:19.686406 10926 solver.cpp:245] Train net output #30: loss1/loss04 = 4.46674 (* 0.0272727 = 0.12182 loss) | |
| I0510 15:32:19.686415 10926 solver.cpp:245] Train net output #31: loss1/loss05 = 3.66893 (* 0.0272727 = 0.100062 loss) | |
| I0510 15:32:19.686424 10926 solver.cpp:245] Train net output #32: loss1/loss06 = 2.93411 (* 0.0272727 = 0.0800212 loss) | |
| I0510 15:32:19.686434 10926 solver.cpp:245] Train net output #33: loss1/loss07 = 1.89014 (* 0.0272727 = 0.0515493 loss) | |
| I0510 15:32:19.686444 10926 solver.cpp:245] Train net output #34: loss1/loss08 = 0.990756 (* 0.0272727 = 0.0270206 loss) | |
| I0510 15:32:19.686452 10926 solver.cpp:245] Train net output #35: loss1/loss09 = 0.129677 (* 0.0272727 = 0.00353664 loss) | |
| I0510 15:32:19.686462 10926 solver.cpp:245] Train net output #36: loss1/loss10 = 0.0986299 (* 0.0272727 = 0.00268991 loss) | |
| I0510 15:32:19.686471 10926 solver.cpp:245] Train net output #37: loss1/loss11 = 0.0950939 (* 0.0272727 = 0.00259347 loss) | |
| I0510 15:32:19.686481 10926 solver.cpp:245] Train net output #38: loss1/loss12 = 0.0686112 (* 0.0272727 = 0.00187121 loss) | |
| I0510 15:32:19.686491 10926 solver.cpp:245] Train net output #39: loss1/loss13 = 0.0601155 (* 0.0272727 = 0.00163951 loss) | |
| I0510 15:32:19.686520 10926 solver.cpp:245] Train net output #40: loss1/loss14 = 0.054337 (* 0.0272727 = 0.00148192 loss) | |
| I0510 15:32:19.686532 10926 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0525692 (* 0.0272727 = 0.00143371 loss) | |
| I0510 15:32:19.686542 10926 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0317427 (* 0.0272727 = 0.00086571 loss) | |
| I0510 15:32:19.686552 10926 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0238156 (* 0.0272727 = 0.000649518 loss) | |
| I0510 15:32:19.686561 10926 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0228264 (* 0.0272727 = 0.000622539 loss) | |
| I0510 15:32:19.686570 10926 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0224233 (* 0.0272727 = 0.000611544 loss) | |
| I0510 15:32:19.686580 10926 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0151623 (* 0.0272727 = 0.000413516 loss) | |
| I0510 15:32:19.686589 10926 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0260937 (* 0.0272727 = 0.000711646 loss) | |
| I0510 15:32:19.686599 10926 solver.cpp:245] Train net output #48: loss1/loss22 = 0.02387 (* 0.0272727 = 0.000650999 loss) | |
| I0510 15:32:19.686607 10926 solver.cpp:245] Train net output #49: loss2/accuracy = 0 | |
| I0510 15:32:19.686614 10926 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0510 15:32:19.686622 10926 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0.125 | |
| I0510 15:32:19.686630 10926 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0510 15:32:19.686637 10926 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.125 | |
| I0510 15:32:19.686645 10926 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.25 | |
| I0510 15:32:19.686652 10926 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.5 | |
| I0510 15:32:19.686660 10926 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.75 | |
| I0510 15:32:19.686667 10926 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.875 | |
| I0510 15:32:19.686674 10926 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0510 15:32:19.686681 10926 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0510 15:32:19.686688 10926 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0510 15:32:19.686697 10926 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0510 15:32:19.686703 10926 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0510 15:32:19.686710 10926 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0510 15:32:19.686717 10926 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0510 15:32:19.686725 10926 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0510 15:32:19.686733 10926 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0510 15:32:19.686739 10926 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0510 15:32:19.686746 10926 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0510 15:32:19.686753 10926 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0510 15:32:19.686760 10926 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0510 15:32:19.686767 10926 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0510 15:32:19.686774 10926 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.75 | |
| I0510 15:32:19.686782 10926 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.0681818 | |
| I0510 15:32:19.686792 10926 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 4.50338 (* 0.3 = 1.35101 loss) | |
| I0510 15:32:19.686801 10926 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.40212 (* 0.3 = 0.420637 loss) | |
| I0510 15:32:19.686811 10926 solver.cpp:245] Train net output #76: loss2/loss01 = 4.12615 (* 0.0272727 = 0.112531 loss) | |
| I0510 15:32:19.686820 10926 solver.cpp:245] Train net output #77: loss2/loss02 = 4.14025 (* 0.0272727 = 0.112916 loss) | |
| I0510 15:32:19.686830 10926 solver.cpp:245] Train net output #78: loss2/loss03 = 4.40525 (* 0.0272727 = 0.120143 loss) | |
| I0510 15:32:19.686848 10926 solver.cpp:245] Train net output #79: loss2/loss04 = 4.11843 (* 0.0272727 = 0.112321 loss) | |
| I0510 15:32:19.686859 10926 solver.cpp:245] Train net output #80: loss2/loss05 = 4.37403 (* 0.0272727 = 0.119292 loss) | |
| I0510 15:32:19.686868 10926 solver.cpp:245] Train net output #81: loss2/loss06 = 3.11353 (* 0.0272727 = 0.0849145 loss) | |
| I0510 15:32:19.686880 10926 solver.cpp:245] Train net output #82: loss2/loss07 = 1.71072 (* 0.0272727 = 0.046656 loss) | |
| I0510 15:32:19.686889 10926 solver.cpp:245] Train net output #83: loss2/loss08 = 0.836485 (* 0.0272727 = 0.0228132 loss) | |
| I0510 15:32:19.686899 10926 solver.cpp:245] Train net output #84: loss2/loss09 = 0.0958161 (* 0.0272727 = 0.00261317 loss) | |
| I0510 15:32:19.686909 10926 solver.cpp:245] Train net output #85: loss2/loss10 = 0.0688084 (* 0.0272727 = 0.00187659 loss) | |
| I0510 15:32:19.686918 10926 solver.cpp:245] Train net output #86: loss2/loss11 = 0.0855367 (* 0.0272727 = 0.00233282 loss) | |
| I0510 15:32:19.686928 10926 solver.cpp:245] Train net output #87: loss2/loss12 = 0.0538168 (* 0.0272727 = 0.00146773 loss) | |
| I0510 15:32:19.686938 10926 solver.cpp:245] Train net output #88: loss2/loss13 = 0.0531541 (* 0.0272727 = 0.00144966 loss) | |
| I0510 15:32:19.686947 10926 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0591991 (* 0.0272727 = 0.00161452 loss) | |
| I0510 15:32:19.686956 10926 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0517758 (* 0.0272727 = 0.00141207 loss) | |
| I0510 15:32:19.686966 10926 solver.cpp:245] Train net output #91: loss2/loss16 = 0.023156 (* 0.0272727 = 0.000631526 loss) | |
| I0510 15:32:19.686975 10926 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0279475 (* 0.0272727 = 0.000762203 loss) | |
| I0510 15:32:19.686985 10926 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0344257 (* 0.0272727 = 0.000938882 loss) | |
| I0510 15:32:19.686995 10926 solver.cpp:245] Train net output #94: loss2/loss19 = 0.039037 (* 0.0272727 = 0.00106465 loss) | |
| I0510 15:32:19.687003 10926 solver.cpp:245] Train net output #95: loss2/loss20 = 0.0241185 (* 0.0272727 = 0.000657776 loss) | |
| I0510 15:32:19.687012 10926 solver.cpp:245] Train net output #96: loss2/loss21 = 0.0355609 (* 0.0272727 = 0.000969844 loss) | |
| I0510 15:32:19.687022 10926 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0144881 (* 0.0272727 = 0.000395131 loss) | |
| I0510 15:32:19.687031 10926 solver.cpp:245] Train net output #98: loss3/accuracy = 0 | |
| I0510 15:32:19.687037 10926 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0510 15:32:19.687046 10926 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0510 15:32:19.687052 10926 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0.125 | |
| I0510 15:32:19.687060 10926 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0510 15:32:19.687067 10926 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.25 | |
| I0510 15:32:19.687075 10926 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.5 | |
| I0510 15:32:19.687083 10926 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.75 | |
| I0510 15:32:19.687089 10926 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.875 | |
| I0510 15:32:19.687098 10926 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0510 15:32:19.687104 10926 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0510 15:32:19.687113 10926 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0510 15:32:19.687119 10926 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0510 15:32:19.687126 10926 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0510 15:32:19.687134 10926 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0510 15:32:19.687140 10926 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0510 15:32:19.687147 10926 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0510 15:32:19.687155 10926 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0510 15:32:19.687175 10926 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0510 15:32:19.687185 10926 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0510 15:32:19.687191 10926 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0510 15:32:19.687199 10926 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0510 15:32:19.687206 10926 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0510 15:32:19.687213 10926 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.721591 | |
| I0510 15:32:19.687222 10926 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.113636 | |
| I0510 15:32:19.687232 10926 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 4.26647 (* 1 = 4.26647 loss) | |
| I0510 15:32:19.687242 10926 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.35968 (* 1 = 1.35968 loss) | |
| I0510 15:32:19.687250 10926 solver.cpp:245] Train net output #125: loss3/loss01 = 4.00235 (* 0.0909091 = 0.36385 loss) | |
| I0510 15:32:19.687259 10926 solver.cpp:245] Train net output #126: loss3/loss02 = 4.06372 (* 0.0909091 = 0.369429 loss) | |
| I0510 15:32:19.687269 10926 solver.cpp:245] Train net output #127: loss3/loss03 = 4.15111 (* 0.0909091 = 0.377374 loss) | |
| I0510 15:32:19.687278 10926 solver.cpp:245] Train net output #128: loss3/loss04 = 3.64179 (* 0.0909091 = 0.331072 loss) | |
| I0510 15:32:19.687288 10926 solver.cpp:245] Train net output #129: loss3/loss05 = 4.24267 (* 0.0909091 = 0.385698 loss) | |
| I0510 15:32:19.687296 10926 solver.cpp:245] Train net output #130: loss3/loss06 = 2.82287 (* 0.0909091 = 0.256625 loss) | |
| I0510 15:32:19.687305 10926 solver.cpp:245] Train net output #131: loss3/loss07 = 1.69388 (* 0.0909091 = 0.153989 loss) | |
| I0510 15:32:19.687314 10926 solver.cpp:245] Train net output #132: loss3/loss08 = 0.782737 (* 0.0909091 = 0.0711579 loss) | |
| I0510 15:32:19.687325 10926 solver.cpp:245] Train net output #133: loss3/loss09 = 0.116484 (* 0.0909091 = 0.0105894 loss) | |
| I0510 15:32:19.687333 10926 solver.cpp:245] Train net output #134: loss3/loss10 = 0.0949613 (* 0.0909091 = 0.00863285 loss) | |
| I0510 15:32:19.687343 10926 solver.cpp:245] Train net output #135: loss3/loss11 = 0.0739903 (* 0.0909091 = 0.00672639 loss) | |
| I0510 15:32:19.687352 10926 solver.cpp:245] Train net output #136: loss3/loss12 = 0.050881 (* 0.0909091 = 0.00462555 loss) | |
| I0510 15:32:19.687361 10926 solver.cpp:245] Train net output #137: loss3/loss13 = 0.0379727 (* 0.0909091 = 0.00345206 loss) | |
| I0510 15:32:19.687371 10926 solver.cpp:245] Train net output #138: loss3/loss14 = 0.0396317 (* 0.0909091 = 0.00360289 loss) | |
| I0510 15:32:19.687379 10926 solver.cpp:245] Train net output #139: loss3/loss15 = 0.0284969 (* 0.0909091 = 0.00259062 loss) | |
| I0510 15:32:19.687389 10926 solver.cpp:245] Train net output #140: loss3/loss16 = 0.0152056 (* 0.0909091 = 0.00138233 loss) | |
| I0510 15:32:19.687398 10926 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00915841 (* 0.0909091 = 0.000832582 loss) | |
| I0510 15:32:19.687407 10926 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00624088 (* 0.0909091 = 0.000567353 loss) | |
| I0510 15:32:19.687417 10926 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00561888 (* 0.0909091 = 0.000510807 loss) | |
| I0510 15:32:19.687427 10926 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00411323 (* 0.0909091 = 0.00037393 loss) | |
| I0510 15:32:19.687435 10926 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00433027 (* 0.0909091 = 0.000393661 loss) | |
| I0510 15:32:19.687444 10926 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00446129 (* 0.0909091 = 0.000405572 loss) | |
| I0510 15:32:19.687453 10926 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0510 15:32:19.687459 10926 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0510 15:32:19.687466 10926 solver.cpp:245] Train net output #149: total_confidence = 4.49923e-09 | |
| I0510 15:32:19.687474 10926 solver.cpp:245] Train net output #150: total_confidence_not_rec = 1.39281e-06 | |
| I0510 15:32:19.687492 10926 sgd_solver.cpp:106] Iteration 1500, lr = 0.001 | |
| I0510 15:34:46.602288 10926 solver.cpp:229] Iteration 2000, loss = 12.8344 | |
| I0510 15:34:46.602421 10926 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0243902 | |
| I0510 15:34:46.602439 10926 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0510 15:34:46.602452 10926 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0510 15:34:46.602465 10926 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0510 15:34:46.602480 10926 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.25 | |
| I0510 15:34:46.602494 10926 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.5 | |
| I0510 15:34:46.602505 10926 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.625 | |
| I0510 15:34:46.602519 10926 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.625 | |
| I0510 15:34:46.602530 10926 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.875 | |
| I0510 15:34:46.602542 10926 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0510 15:34:46.602555 10926 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0510 15:34:46.602567 10926 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0510 15:34:46.602579 10926 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0510 15:34:46.602591 10926 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0510 15:34:46.602602 10926 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0510 15:34:46.602614 10926 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0510 15:34:46.602627 10926 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0510 15:34:46.602638 10926 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0510 15:34:46.602649 10926 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0510 15:34:46.602661 10926 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0510 15:34:46.602672 10926 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0510 15:34:46.602684 10926 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0510 15:34:46.602695 10926 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0510 15:34:46.602707 10926 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.772727 | |
| I0510 15:34:46.602718 10926 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.097561 | |
| I0510 15:34:46.602738 10926 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.96984 (* 0.3 = 1.19095 loss) | |
| I0510 15:34:46.602753 10926 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.22684 (* 0.3 = 0.368051 loss) | |
| I0510 15:34:46.602768 10926 solver.cpp:245] Train net output #27: loss1/loss01 = 4.04192 (* 0.0272727 = 0.110234 loss) | |
| I0510 15:34:46.602782 10926 solver.cpp:245] Train net output #28: loss1/loss02 = 3.80638 (* 0.0272727 = 0.10381 loss) | |
| I0510 15:34:46.602797 10926 solver.cpp:245] Train net output #29: loss1/loss03 = 3.95894 (* 0.0272727 = 0.107971 loss) | |
| I0510 15:34:46.602812 10926 solver.cpp:245] Train net output #30: loss1/loss04 = 3.55109 (* 0.0272727 = 0.0968478 loss) | |
| I0510 15:34:46.602825 10926 solver.cpp:245] Train net output #31: loss1/loss05 = 2.94485 (* 0.0272727 = 0.080314 loss) | |
| I0510 15:34:46.602839 10926 solver.cpp:245] Train net output #32: loss1/loss06 = 2.44053 (* 0.0272727 = 0.0665599 loss) | |
| I0510 15:34:46.602854 10926 solver.cpp:245] Train net output #33: loss1/loss07 = 2.62343 (* 0.0272727 = 0.071548 loss) | |
| I0510 15:34:46.602867 10926 solver.cpp:245] Train net output #34: loss1/loss08 = 0.863473 (* 0.0272727 = 0.0235493 loss) | |
| I0510 15:34:46.602882 10926 solver.cpp:245] Train net output #35: loss1/loss09 = 0.136048 (* 0.0272727 = 0.00371039 loss) | |
| I0510 15:34:46.602897 10926 solver.cpp:245] Train net output #36: loss1/loss10 = 0.121801 (* 0.0272727 = 0.00332185 loss) | |
| I0510 15:34:46.602911 10926 solver.cpp:245] Train net output #37: loss1/loss11 = 0.122596 (* 0.0272727 = 0.00334351 loss) | |
| I0510 15:34:46.602926 10926 solver.cpp:245] Train net output #38: loss1/loss12 = 0.0596803 (* 0.0272727 = 0.00162765 loss) | |
| I0510 15:34:46.602941 10926 solver.cpp:245] Train net output #39: loss1/loss13 = 0.0721858 (* 0.0272727 = 0.0019687 loss) | |
| I0510 15:34:46.602972 10926 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0390978 (* 0.0272727 = 0.0010663 loss) | |
| I0510 15:34:46.602988 10926 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0530767 (* 0.0272727 = 0.00144755 loss) | |
| I0510 15:34:46.603003 10926 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0392361 (* 0.0272727 = 0.00107007 loss) | |
| I0510 15:34:46.603018 10926 solver.cpp:245] Train net output #43: loss1/loss17 = 0.035712 (* 0.0272727 = 0.000973963 loss) | |
| I0510 15:34:46.603032 10926 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0253581 (* 0.0272727 = 0.000691586 loss) | |
| I0510 15:34:46.603046 10926 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0292026 (* 0.0272727 = 0.000796436 loss) | |
| I0510 15:34:46.603060 10926 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0362884 (* 0.0272727 = 0.000989684 loss) | |
| I0510 15:34:46.603075 10926 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0292125 (* 0.0272727 = 0.000796706 loss) | |
| I0510 15:34:46.603090 10926 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0253237 (* 0.0272727 = 0.000690646 loss) | |
| I0510 15:34:46.603101 10926 solver.cpp:245] Train net output #49: loss2/accuracy = 0 | |
| I0510 15:34:46.603113 10926 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.125 | |
| I0510 15:34:46.603126 10926 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0510 15:34:46.603137 10926 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0.125 | |
| I0510 15:34:46.603149 10926 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.125 | |
| I0510 15:34:46.603160 10926 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.5 | |
| I0510 15:34:46.603173 10926 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.625 | |
| I0510 15:34:46.603185 10926 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.625 | |
| I0510 15:34:46.603196 10926 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.875 | |
| I0510 15:34:46.603209 10926 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0510 15:34:46.603220 10926 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0510 15:34:46.603231 10926 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0510 15:34:46.603243 10926 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0510 15:34:46.603255 10926 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0510 15:34:46.603266 10926 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0510 15:34:46.603277 10926 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0510 15:34:46.603289 10926 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0510 15:34:46.603301 10926 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0510 15:34:46.603312 10926 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0510 15:34:46.603323 10926 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0510 15:34:46.603335 10926 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0510 15:34:46.603346 10926 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0510 15:34:46.603358 10926 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0510 15:34:46.603369 10926 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.767045 | |
| I0510 15:34:46.603381 10926 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.0487805 | |
| I0510 15:34:46.603395 10926 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.87656 (* 0.3 = 1.16297 loss) | |
| I0510 15:34:46.603410 10926 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.2316 (* 0.3 = 0.369479 loss) | |
| I0510 15:34:46.603423 10926 solver.cpp:245] Train net output #76: loss2/loss01 = 3.79337 (* 0.0272727 = 0.103456 loss) | |
| I0510 15:34:46.603437 10926 solver.cpp:245] Train net output #77: loss2/loss02 = 3.89688 (* 0.0272727 = 0.106279 loss) | |
| I0510 15:34:46.603452 10926 solver.cpp:245] Train net output #78: loss2/loss03 = 3.67114 (* 0.0272727 = 0.100122 loss) | |
| I0510 15:34:46.603477 10926 solver.cpp:245] Train net output #79: loss2/loss04 = 3.65073 (* 0.0272727 = 0.0995653 loss) | |
| I0510 15:34:46.603493 10926 solver.cpp:245] Train net output #80: loss2/loss05 = 3.30411 (* 0.0272727 = 0.0901122 loss) | |
| I0510 15:34:46.603507 10926 solver.cpp:245] Train net output #81: loss2/loss06 = 2.35596 (* 0.0272727 = 0.0642535 loss) | |
| I0510 15:34:46.603521 10926 solver.cpp:245] Train net output #82: loss2/loss07 = 2.28294 (* 0.0272727 = 0.062262 loss) | |
| I0510 15:34:46.603539 10926 solver.cpp:245] Train net output #83: loss2/loss08 = 0.904007 (* 0.0272727 = 0.0246547 loss) | |
| I0510 15:34:46.603554 10926 solver.cpp:245] Train net output #84: loss2/loss09 = 0.287673 (* 0.0272727 = 0.00784562 loss) | |
| I0510 15:34:46.603567 10926 solver.cpp:245] Train net output #85: loss2/loss10 = 0.11345 (* 0.0272727 = 0.00309409 loss) | |
| I0510 15:34:46.603582 10926 solver.cpp:245] Train net output #86: loss2/loss11 = 0.115485 (* 0.0272727 = 0.0031496 loss) | |
| I0510 15:34:46.603596 10926 solver.cpp:245] Train net output #87: loss2/loss12 = 0.112879 (* 0.0272727 = 0.00307851 loss) | |
| I0510 15:34:46.603610 10926 solver.cpp:245] Train net output #88: loss2/loss13 = 0.0595575 (* 0.0272727 = 0.0016243 loss) | |
| I0510 15:34:46.603624 10926 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0628577 (* 0.0272727 = 0.0017143 loss) | |
| I0510 15:34:46.603639 10926 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0984429 (* 0.0272727 = 0.00268481 loss) | |
| I0510 15:34:46.603653 10926 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0571044 (* 0.0272727 = 0.00155739 loss) | |
| I0510 15:34:46.603667 10926 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0305859 (* 0.0272727 = 0.000834161 loss) | |
| I0510 15:34:46.603682 10926 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0388879 (* 0.0272727 = 0.00106058 loss) | |
| I0510 15:34:46.603695 10926 solver.cpp:245] Train net output #94: loss2/loss19 = 0.0391766 (* 0.0272727 = 0.00106845 loss) | |
| I0510 15:34:46.603708 10926 solver.cpp:245] Train net output #95: loss2/loss20 = 0.0359553 (* 0.0272727 = 0.000980599 loss) | |
| I0510 15:34:46.603724 10926 solver.cpp:245] Train net output #96: loss2/loss21 = 0.0334029 (* 0.0272727 = 0.000910989 loss) | |
| I0510 15:34:46.603737 10926 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0489781 (* 0.0272727 = 0.00133577 loss) | |
| I0510 15:34:46.603749 10926 solver.cpp:245] Train net output #98: loss3/accuracy = 0 | |
| I0510 15:34:46.603761 10926 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.125 | |
| I0510 15:34:46.603773 10926 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0510 15:34:46.603787 10926 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0510 15:34:46.603799 10926 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.25 | |
| I0510 15:34:46.603811 10926 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.5 | |
| I0510 15:34:46.603823 10926 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.625 | |
| I0510 15:34:46.603835 10926 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.625 | |
| I0510 15:34:46.603847 10926 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.875 | |
| I0510 15:34:46.603857 10926 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0510 15:34:46.603869 10926 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0510 15:34:46.603880 10926 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0510 15:34:46.603893 10926 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0510 15:34:46.603904 10926 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0510 15:34:46.603914 10926 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0510 15:34:46.603925 10926 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0510 15:34:46.603937 10926 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0510 15:34:46.603948 10926 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0510 15:34:46.603971 10926 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0510 15:34:46.603983 10926 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0510 15:34:46.603994 10926 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0510 15:34:46.604007 10926 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0510 15:34:46.604018 10926 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0510 15:34:46.604029 10926 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.75 | |
| I0510 15:34:46.604043 10926 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.097561 | |
| I0510 15:34:46.604056 10926 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.79813 (* 1 = 3.79813 loss) | |
| I0510 15:34:46.604070 10926 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.13835 (* 1 = 1.13835 loss) | |
| I0510 15:34:46.604084 10926 solver.cpp:245] Train net output #125: loss3/loss01 = 3.81225 (* 0.0909091 = 0.346568 loss) | |
| I0510 15:34:46.604099 10926 solver.cpp:245] Train net output #126: loss3/loss02 = 3.43704 (* 0.0909091 = 0.312458 loss) | |
| I0510 15:34:46.604112 10926 solver.cpp:245] Train net output #127: loss3/loss03 = 3.62428 (* 0.0909091 = 0.32948 loss) | |
| I0510 15:34:46.604126 10926 solver.cpp:245] Train net output #128: loss3/loss04 = 3.6802 (* 0.0909091 = 0.334563 loss) | |
| I0510 15:34:46.604140 10926 solver.cpp:245] Train net output #129: loss3/loss05 = 2.59852 (* 0.0909091 = 0.236229 loss) | |
| I0510 15:34:46.604153 10926 solver.cpp:245] Train net output #130: loss3/loss06 = 2.25258 (* 0.0909091 = 0.20478 loss) | |
| I0510 15:34:46.604167 10926 solver.cpp:245] Train net output #131: loss3/loss07 = 2.07905 (* 0.0909091 = 0.189004 loss) | |
| I0510 15:34:46.604182 10926 solver.cpp:245] Train net output #132: loss3/loss08 = 1.01315 (* 0.0909091 = 0.0921046 loss) | |
| I0510 15:34:46.604192 10926 solver.cpp:245] Train net output #133: loss3/loss09 = 0.154111 (* 0.0909091 = 0.0140101 loss) | |
| I0510 15:34:46.604207 10926 solver.cpp:245] Train net output #134: loss3/loss10 = 0.125514 (* 0.0909091 = 0.0114104 loss) | |
| I0510 15:34:46.604220 10926 solver.cpp:245] Train net output #135: loss3/loss11 = 0.0851708 (* 0.0909091 = 0.0077428 loss) | |
| I0510 15:34:46.604234 10926 solver.cpp:245] Train net output #136: loss3/loss12 = 0.0695633 (* 0.0909091 = 0.00632394 loss) | |
| I0510 15:34:46.604249 10926 solver.cpp:245] Train net output #137: loss3/loss13 = 0.0560731 (* 0.0909091 = 0.00509756 loss) | |
| I0510 15:34:46.604262 10926 solver.cpp:245] Train net output #138: loss3/loss14 = 0.0363723 (* 0.0909091 = 0.00330657 loss) | |
| I0510 15:34:46.604276 10926 solver.cpp:245] Train net output #139: loss3/loss15 = 0.0399042 (* 0.0909091 = 0.00362765 loss) | |
| I0510 15:34:46.604290 10926 solver.cpp:245] Train net output #140: loss3/loss16 = 0.0229927 (* 0.0909091 = 0.00209025 loss) | |
| I0510 15:34:46.604303 10926 solver.cpp:245] Train net output #141: loss3/loss17 = 0.0166686 (* 0.0909091 = 0.00151533 loss) | |
| I0510 15:34:46.604317 10926 solver.cpp:245] Train net output #142: loss3/loss18 = 0.0167551 (* 0.0909091 = 0.00152319 loss) | |
| I0510 15:34:46.604331 10926 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00995279 (* 0.0909091 = 0.0009048 loss) | |
| I0510 15:34:46.604346 10926 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00938336 (* 0.0909091 = 0.000853033 loss) | |
| I0510 15:34:46.604359 10926 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00808689 (* 0.0909091 = 0.000735172 loss) | |
| I0510 15:34:46.604373 10926 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00695948 (* 0.0909091 = 0.00063268 loss) | |
| I0510 15:34:46.604385 10926 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0510 15:34:46.604396 10926 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0510 15:34:46.604408 10926 solver.cpp:245] Train net output #149: total_confidence = 1.30567e-08 | |
| I0510 15:34:46.604419 10926 solver.cpp:245] Train net output #150: total_confidence_not_rec = 1.25804e-06 | |
| I0510 15:34:46.604442 10926 sgd_solver.cpp:106] Iteration 2000, lr = 0.001 | |
| I0510 15:37:13.428721 10926 solver.cpp:229] Iteration 2500, loss = 12.7503 | |
| I0510 15:37:13.428961 10926 solver.cpp:245] Train net output #0: loss1/accuracy = 0 | |
| I0510 15:37:13.428982 10926 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.125 | |
| I0510 15:37:13.428997 10926 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0510 15:37:13.429008 10926 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0510 15:37:13.429020 10926 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.125 | |
| I0510 15:37:13.429033 10926 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.125 | |
| I0510 15:37:13.429044 10926 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.125 | |
| I0510 15:37:13.429056 10926 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.25 | |
| I0510 15:37:13.429067 10926 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.625 | |
| I0510 15:37:13.429080 10926 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.75 | |
| I0510 15:37:13.429091 10926 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.875 | |
| I0510 15:37:13.429105 10926 solver.cpp:245] Train net output #11: loss1/accuracy11 = 0.875 | |
| I0510 15:37:13.429131 10926 solver.cpp:245] Train net output #12: loss1/accuracy12 = 0.875 | |
| I0510 15:37:13.429147 10926 solver.cpp:245] Train net output #13: loss1/accuracy13 = 0.875 | |
| I0510 15:37:13.429160 10926 solver.cpp:245] Train net output #14: loss1/accuracy14 = 0.875 | |
| I0510 15:37:13.429172 10926 solver.cpp:245] Train net output #15: loss1/accuracy15 = 0.875 | |
| I0510 15:37:13.429184 10926 solver.cpp:245] Train net output #16: loss1/accuracy16 = 0.875 | |
| I0510 15:37:13.429195 10926 solver.cpp:245] Train net output #17: loss1/accuracy17 = 0.875 | |
| I0510 15:37:13.429208 10926 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0510 15:37:13.429226 10926 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0510 15:37:13.429239 10926 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0510 15:37:13.429250 10926 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0510 15:37:13.429261 10926 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0510 15:37:13.429273 10926 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.636364 | |
| I0510 15:37:13.429286 10926 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.09375 | |
| I0510 15:37:13.429301 10926 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 4.13882 (* 0.3 = 1.24165 loss) | |
| I0510 15:37:13.429316 10926 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.65875 (* 0.3 = 0.497624 loss) | |
| I0510 15:37:13.429330 10926 solver.cpp:245] Train net output #27: loss1/loss01 = 3.90623 (* 0.0272727 = 0.106534 loss) | |
| I0510 15:37:13.429345 10926 solver.cpp:245] Train net output #28: loss1/loss02 = 4.29413 (* 0.0272727 = 0.117113 loss) | |
| I0510 15:37:13.429359 10926 solver.cpp:245] Train net output #29: loss1/loss03 = 3.76755 (* 0.0272727 = 0.102751 loss) | |
| I0510 15:37:13.429373 10926 solver.cpp:245] Train net output #30: loss1/loss04 = 3.72633 (* 0.0272727 = 0.101627 loss) | |
| I0510 15:37:13.429388 10926 solver.cpp:245] Train net output #31: loss1/loss05 = 3.56051 (* 0.0272727 = 0.0971047 loss) | |
| I0510 15:37:13.429402 10926 solver.cpp:245] Train net output #32: loss1/loss06 = 3.8146 (* 0.0272727 = 0.104035 loss) | |
| I0510 15:37:13.429416 10926 solver.cpp:245] Train net output #33: loss1/loss07 = 3.61326 (* 0.0272727 = 0.0985435 loss) | |
| I0510 15:37:13.429430 10926 solver.cpp:245] Train net output #34: loss1/loss08 = 2.03904 (* 0.0272727 = 0.0556103 loss) | |
| I0510 15:37:13.429443 10926 solver.cpp:245] Train net output #35: loss1/loss09 = 1.41538 (* 0.0272727 = 0.0386014 loss) | |
| I0510 15:37:13.429458 10926 solver.cpp:245] Train net output #36: loss1/loss10 = 0.775032 (* 0.0272727 = 0.0211372 loss) | |
| I0510 15:37:13.429472 10926 solver.cpp:245] Train net output #37: loss1/loss11 = 0.866621 (* 0.0272727 = 0.0236351 loss) | |
| I0510 15:37:13.429486 10926 solver.cpp:245] Train net output #38: loss1/loss12 = 0.786675 (* 0.0272727 = 0.0214548 loss) | |
| I0510 15:37:13.429518 10926 solver.cpp:245] Train net output #39: loss1/loss13 = 0.707292 (* 0.0272727 = 0.0192898 loss) | |
| I0510 15:37:13.429534 10926 solver.cpp:245] Train net output #40: loss1/loss14 = 0.629898 (* 0.0272727 = 0.017179 loss) | |
| I0510 15:37:13.429548 10926 solver.cpp:245] Train net output #41: loss1/loss15 = 0.94035 (* 0.0272727 = 0.0256459 loss) | |
| I0510 15:37:13.429563 10926 solver.cpp:245] Train net output #42: loss1/loss16 = 0.918848 (* 0.0272727 = 0.0250595 loss) | |
| I0510 15:37:13.429577 10926 solver.cpp:245] Train net output #43: loss1/loss17 = 0.798596 (* 0.0272727 = 0.0217799 loss) | |
| I0510 15:37:13.429592 10926 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0297335 (* 0.0272727 = 0.000810913 loss) | |
| I0510 15:37:13.429606 10926 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0326721 (* 0.0272727 = 0.000891058 loss) | |
| I0510 15:37:13.429620 10926 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0363038 (* 0.0272727 = 0.000990104 loss) | |
| I0510 15:37:13.429635 10926 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0254669 (* 0.0272727 = 0.000694552 loss) | |
| I0510 15:37:13.429648 10926 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0462812 (* 0.0272727 = 0.00126221 loss) | |
| I0510 15:37:13.429661 10926 solver.cpp:245] Train net output #49: loss2/accuracy = 0.03125 | |
| I0510 15:37:13.429673 10926 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0510 15:37:13.429685 10926 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0510 15:37:13.429697 10926 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0.125 | |
| I0510 15:37:13.429708 10926 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.125 | |
| I0510 15:37:13.429720 10926 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.125 | |
| I0510 15:37:13.429733 10926 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.125 | |
| I0510 15:37:13.429744 10926 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.25 | |
| I0510 15:37:13.429755 10926 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.625 | |
| I0510 15:37:13.429767 10926 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.75 | |
| I0510 15:37:13.429780 10926 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.875 | |
| I0510 15:37:13.429791 10926 solver.cpp:245] Train net output #60: loss2/accuracy11 = 0.875 | |
| I0510 15:37:13.429803 10926 solver.cpp:245] Train net output #61: loss2/accuracy12 = 0.875 | |
| I0510 15:37:13.429814 10926 solver.cpp:245] Train net output #62: loss2/accuracy13 = 0.875 | |
| I0510 15:37:13.429826 10926 solver.cpp:245] Train net output #63: loss2/accuracy14 = 0.875 | |
| I0510 15:37:13.429838 10926 solver.cpp:245] Train net output #64: loss2/accuracy15 = 0.875 | |
| I0510 15:37:13.429850 10926 solver.cpp:245] Train net output #65: loss2/accuracy16 = 0.875 | |
| I0510 15:37:13.429862 10926 solver.cpp:245] Train net output #66: loss2/accuracy17 = 0.875 | |
| I0510 15:37:13.429877 10926 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0510 15:37:13.429888 10926 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0510 15:37:13.429900 10926 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0510 15:37:13.429913 10926 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0510 15:37:13.429924 10926 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0510 15:37:13.429934 10926 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.647727 | |
| I0510 15:37:13.429946 10926 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.15625 | |
| I0510 15:37:13.429960 10926 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 4.18044 (* 0.3 = 1.25413 loss) | |
| I0510 15:37:13.429975 10926 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.69959 (* 0.3 = 0.509876 loss) | |
| I0510 15:37:13.429991 10926 solver.cpp:245] Train net output #76: loss2/loss01 = 3.92618 (* 0.0272727 = 0.107078 loss) | |
| I0510 15:37:13.430006 10926 solver.cpp:245] Train net output #77: loss2/loss02 = 4.08049 (* 0.0272727 = 0.111286 loss) | |
| I0510 15:37:13.430032 10926 solver.cpp:245] Train net output #78: loss2/loss03 = 3.79967 (* 0.0272727 = 0.103627 loss) | |
| I0510 15:37:13.430047 10926 solver.cpp:245] Train net output #79: loss2/loss04 = 3.77824 (* 0.0272727 = 0.103043 loss) | |
| I0510 15:37:13.430060 10926 solver.cpp:245] Train net output #80: loss2/loss05 = 3.94476 (* 0.0272727 = 0.107584 loss) | |
| I0510 15:37:13.430074 10926 solver.cpp:245] Train net output #81: loss2/loss06 = 4.06844 (* 0.0272727 = 0.110957 loss) | |
| I0510 15:37:13.430088 10926 solver.cpp:245] Train net output #82: loss2/loss07 = 3.84383 (* 0.0272727 = 0.104832 loss) | |
| I0510 15:37:13.430101 10926 solver.cpp:245] Train net output #83: loss2/loss08 = 2.05928 (* 0.0272727 = 0.0561622 loss) | |
| I0510 15:37:13.430115 10926 solver.cpp:245] Train net output #84: loss2/loss09 = 1.72694 (* 0.0272727 = 0.0470985 loss) | |
| I0510 15:37:13.430130 10926 solver.cpp:245] Train net output #85: loss2/loss10 = 0.857286 (* 0.0272727 = 0.0233805 loss) | |
| I0510 15:37:13.430143 10926 solver.cpp:245] Train net output #86: loss2/loss11 = 0.738494 (* 0.0272727 = 0.0201408 loss) | |
| I0510 15:37:13.430157 10926 solver.cpp:245] Train net output #87: loss2/loss12 = 0.977967 (* 0.0272727 = 0.0266718 loss) | |
| I0510 15:37:13.430171 10926 solver.cpp:245] Train net output #88: loss2/loss13 = 0.837278 (* 0.0272727 = 0.0228348 loss) | |
| I0510 15:37:13.430186 10926 solver.cpp:245] Train net output #89: loss2/loss14 = 0.869885 (* 0.0272727 = 0.0237241 loss) | |
| I0510 15:37:13.430199 10926 solver.cpp:245] Train net output #90: loss2/loss15 = 0.820069 (* 0.0272727 = 0.0223655 loss) | |
| I0510 15:37:13.430213 10926 solver.cpp:245] Train net output #91: loss2/loss16 = 0.803092 (* 0.0272727 = 0.0219025 loss) | |
| I0510 15:37:13.430227 10926 solver.cpp:245] Train net output #92: loss2/loss17 = 0.835636 (* 0.0272727 = 0.0227901 loss) | |
| I0510 15:37:13.430243 10926 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0333477 (* 0.0272727 = 0.000909484 loss) | |
| I0510 15:37:13.430256 10926 solver.cpp:245] Train net output #94: loss2/loss19 = 0.0397274 (* 0.0272727 = 0.00108348 loss) | |
| I0510 15:37:13.430270 10926 solver.cpp:245] Train net output #95: loss2/loss20 = 0.0353865 (* 0.0272727 = 0.000965087 loss) | |
| I0510 15:37:13.430284 10926 solver.cpp:245] Train net output #96: loss2/loss21 = 0.0278804 (* 0.0272727 = 0.000760375 loss) | |
| I0510 15:37:13.430299 10926 solver.cpp:245] Train net output #97: loss2/loss22 = 0.025024 (* 0.0272727 = 0.000682472 loss) | |
| I0510 15:37:13.430310 10926 solver.cpp:245] Train net output #98: loss3/accuracy = 0 | |
| I0510 15:37:13.430322 10926 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0510 15:37:13.430335 10926 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0510 15:37:13.430342 10926 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0.125 | |
| I0510 15:37:13.430351 10926 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0510 15:37:13.430362 10926 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.125 | |
| I0510 15:37:13.430374 10926 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.125 | |
| I0510 15:37:13.430387 10926 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.25 | |
| I0510 15:37:13.430397 10926 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.625 | |
| I0510 15:37:13.430409 10926 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.75 | |
| I0510 15:37:13.430421 10926 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.875 | |
| I0510 15:37:13.430433 10926 solver.cpp:245] Train net output #109: loss3/accuracy11 = 0.875 | |
| I0510 15:37:13.430444 10926 solver.cpp:245] Train net output #110: loss3/accuracy12 = 0.875 | |
| I0510 15:37:13.430455 10926 solver.cpp:245] Train net output #111: loss3/accuracy13 = 0.875 | |
| I0510 15:37:13.430467 10926 solver.cpp:245] Train net output #112: loss3/accuracy14 = 0.875 | |
| I0510 15:37:13.430479 10926 solver.cpp:245] Train net output #113: loss3/accuracy15 = 0.875 | |
| I0510 15:37:13.430490 10926 solver.cpp:245] Train net output #114: loss3/accuracy16 = 0.875 | |
| I0510 15:37:13.430511 10926 solver.cpp:245] Train net output #115: loss3/accuracy17 = 0.875 | |
| I0510 15:37:13.430526 10926 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0510 15:37:13.430537 10926 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0510 15:37:13.430548 10926 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0510 15:37:13.430559 10926 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0510 15:37:13.430572 10926 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0510 15:37:13.430583 10926 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.630682 | |
| I0510 15:37:13.430594 10926 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.15625 | |
| I0510 15:37:13.430608 10926 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 4.03503 (* 1 = 4.03503 loss) | |
| I0510 15:37:13.430622 10926 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.62739 (* 1 = 1.62739 loss) | |
| I0510 15:37:13.430636 10926 solver.cpp:245] Train net output #125: loss3/loss01 = 3.69041 (* 0.0909091 = 0.335492 loss) | |
| I0510 15:37:13.430650 10926 solver.cpp:245] Train net output #126: loss3/loss02 = 3.71538 (* 0.0909091 = 0.337762 loss) | |
| I0510 15:37:13.430665 10926 solver.cpp:245] Train net output #127: loss3/loss03 = 3.71023 (* 0.0909091 = 0.337294 loss) | |
| I0510 15:37:13.430678 10926 solver.cpp:245] Train net output #128: loss3/loss04 = 3.66991 (* 0.0909091 = 0.333628 loss) | |
| I0510 15:37:13.430692 10926 solver.cpp:245] Train net output #129: loss3/loss05 = 3.55289 (* 0.0909091 = 0.32299 loss) | |
| I0510 15:37:13.430706 10926 solver.cpp:245] Train net output #130: loss3/loss06 = 4.01377 (* 0.0909091 = 0.364888 loss) | |
| I0510 15:37:13.430719 10926 solver.cpp:245] Train net output #131: loss3/loss07 = 3.44434 (* 0.0909091 = 0.313122 loss) | |
| I0510 15:37:13.430733 10926 solver.cpp:245] Train net output #132: loss3/loss08 = 1.89368 (* 0.0909091 = 0.172153 loss) | |
| I0510 15:37:13.430747 10926 solver.cpp:245] Train net output #133: loss3/loss09 = 1.4768 (* 0.0909091 = 0.134254 loss) | |
| I0510 15:37:13.430762 10926 solver.cpp:245] Train net output #134: loss3/loss10 = 0.90876 (* 0.0909091 = 0.0826146 loss) | |
| I0510 15:37:13.430774 10926 solver.cpp:245] Train net output #135: loss3/loss11 = 0.736711 (* 0.0909091 = 0.0669737 loss) | |
| I0510 15:37:13.430788 10926 solver.cpp:245] Train net output #136: loss3/loss12 = 0.970628 (* 0.0909091 = 0.0882389 loss) | |
| I0510 15:37:13.430802 10926 solver.cpp:245] Train net output #137: loss3/loss13 = 0.793373 (* 0.0909091 = 0.0721248 loss) | |
| I0510 15:37:13.430816 10926 solver.cpp:245] Train net output #138: loss3/loss14 = 1.13571 (* 0.0909091 = 0.103246 loss) | |
| I0510 15:37:13.430830 10926 solver.cpp:245] Train net output #139: loss3/loss15 = 0.959642 (* 0.0909091 = 0.0872402 loss) | |
| I0510 15:37:13.430843 10926 solver.cpp:245] Train net output #140: loss3/loss16 = 1.04233 (* 0.0909091 = 0.0947573 loss) | |
| I0510 15:37:13.430857 10926 solver.cpp:245] Train net output #141: loss3/loss17 = 1.12956 (* 0.0909091 = 0.102688 loss) | |
| I0510 15:37:13.430871 10926 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00598276 (* 0.0909091 = 0.000543888 loss) | |
| I0510 15:37:13.430886 10926 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00488528 (* 0.0909091 = 0.000444116 loss) | |
| I0510 15:37:13.430899 10926 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00361566 (* 0.0909091 = 0.000328696 loss) | |
| I0510 15:37:13.430913 10926 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00311458 (* 0.0909091 = 0.000283143 loss) | |
| I0510 15:37:13.430929 10926 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00444051 (* 0.0909091 = 0.000403683 loss) | |
| I0510 15:37:13.430943 10926 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0510 15:37:13.430954 10926 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0510 15:37:13.430966 10926 solver.cpp:245] Train net output #149: total_confidence = 1.18105e-09 | |
| I0510 15:37:13.430989 10926 solver.cpp:245] Train net output #150: total_confidence_not_rec = 2.26632e-06 | |
| I0510 15:37:13.431005 10926 sgd_solver.cpp:106] Iteration 2500, lr = 0.001 | |
| I0510 15:39:40.151521 10926 solver.cpp:229] Iteration 3000, loss = 12.6666 | |
| I0510 15:39:40.151711 10926 solver.cpp:245] Train net output #0: loss1/accuracy = 0 | |
| I0510 15:39:40.151732 10926 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.125 | |
| I0510 15:39:40.151746 10926 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0510 15:39:40.151759 10926 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0510 15:39:40.151772 10926 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.125 | |
| I0510 15:39:40.151783 10926 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.125 | |
| I0510 15:39:40.151795 10926 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.375 | |
| I0510 15:39:40.151808 10926 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.875 | |
| I0510 15:39:40.151820 10926 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.875 | |
| I0510 15:39:40.151832 10926 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0510 15:39:40.151844 10926 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0510 15:39:40.151855 10926 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0510 15:39:40.151867 10926 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0510 15:39:40.151881 10926 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0510 15:39:40.151893 10926 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0510 15:39:40.151904 10926 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0510 15:39:40.151916 10926 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0510 15:39:40.151927 10926 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0510 15:39:40.151939 10926 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0510 15:39:40.151952 10926 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0510 15:39:40.151962 10926 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0510 15:39:40.151974 10926 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0510 15:39:40.151985 10926 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0510 15:39:40.151998 10926 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.744318 | |
| I0510 15:39:40.152009 10926 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.111111 | |
| I0510 15:39:40.152026 10926 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.83495 (* 0.3 = 1.15049 loss) | |
| I0510 15:39:40.152040 10926 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.15035 (* 0.3 = 0.345104 loss) | |
| I0510 15:39:40.152055 10926 solver.cpp:245] Train net output #27: loss1/loss01 = 3.63221 (* 0.0272727 = 0.0990602 loss) | |
| I0510 15:39:40.152070 10926 solver.cpp:245] Train net output #28: loss1/loss02 = 3.84104 (* 0.0272727 = 0.104756 loss) | |
| I0510 15:39:40.152083 10926 solver.cpp:245] Train net output #29: loss1/loss03 = 3.7675 (* 0.0272727 = 0.10275 loss) | |
| I0510 15:39:40.152097 10926 solver.cpp:245] Train net output #30: loss1/loss04 = 4.072 (* 0.0272727 = 0.111055 loss) | |
| I0510 15:39:40.152112 10926 solver.cpp:245] Train net output #31: loss1/loss05 = 3.18479 (* 0.0272727 = 0.086858 loss) | |
| I0510 15:39:40.152127 10926 solver.cpp:245] Train net output #32: loss1/loss06 = 2.87535 (* 0.0272727 = 0.0784187 loss) | |
| I0510 15:39:40.152140 10926 solver.cpp:245] Train net output #33: loss1/loss07 = 1.21999 (* 0.0272727 = 0.0332724 loss) | |
| I0510 15:39:40.152154 10926 solver.cpp:245] Train net output #34: loss1/loss08 = 0.941202 (* 0.0272727 = 0.0256691 loss) | |
| I0510 15:39:40.152168 10926 solver.cpp:245] Train net output #35: loss1/loss09 = 0.201192 (* 0.0272727 = 0.00548705 loss) | |
| I0510 15:39:40.152182 10926 solver.cpp:245] Train net output #36: loss1/loss10 = 0.198646 (* 0.0272727 = 0.00541763 loss) | |
| I0510 15:39:40.152196 10926 solver.cpp:245] Train net output #37: loss1/loss11 = 0.210618 (* 0.0272727 = 0.00574414 loss) | |
| I0510 15:39:40.152211 10926 solver.cpp:245] Train net output #38: loss1/loss12 = 0.127823 (* 0.0272727 = 0.00348608 loss) | |
| I0510 15:39:40.152225 10926 solver.cpp:245] Train net output #39: loss1/loss13 = 0.115014 (* 0.0272727 = 0.00313674 loss) | |
| I0510 15:39:40.152254 10926 solver.cpp:245] Train net output #40: loss1/loss14 = 0.114129 (* 0.0272727 = 0.00311261 loss) | |
| I0510 15:39:40.152269 10926 solver.cpp:245] Train net output #41: loss1/loss15 = 0.113463 (* 0.0272727 = 0.00309445 loss) | |
| I0510 15:39:40.152283 10926 solver.cpp:245] Train net output #42: loss1/loss16 = 0.093349 (* 0.0272727 = 0.00254588 loss) | |
| I0510 15:39:40.152298 10926 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0560314 (* 0.0272727 = 0.00152813 loss) | |
| I0510 15:39:40.152312 10926 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0534249 (* 0.0272727 = 0.00145704 loss) | |
| I0510 15:39:40.152326 10926 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0666215 (* 0.0272727 = 0.00181695 loss) | |
| I0510 15:39:40.152340 10926 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0433302 (* 0.0272727 = 0.00118173 loss) | |
| I0510 15:39:40.152354 10926 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0396849 (* 0.0272727 = 0.00108231 loss) | |
| I0510 15:39:40.152369 10926 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0394323 (* 0.0272727 = 0.00107543 loss) | |
| I0510 15:39:40.152380 10926 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0444444 | |
| I0510 15:39:40.152393 10926 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.125 | |
| I0510 15:39:40.152406 10926 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0.25 | |
| I0510 15:39:40.152418 10926 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0510 15:39:40.152431 10926 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.125 | |
| I0510 15:39:40.152441 10926 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.125 | |
| I0510 15:39:40.152453 10926 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.375 | |
| I0510 15:39:40.152465 10926 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.875 | |
| I0510 15:39:40.152477 10926 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.875 | |
| I0510 15:39:40.152488 10926 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0510 15:39:40.152499 10926 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0510 15:39:40.152511 10926 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0510 15:39:40.152523 10926 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0510 15:39:40.152534 10926 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0510 15:39:40.152545 10926 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0510 15:39:40.152556 10926 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0510 15:39:40.152567 10926 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0510 15:39:40.152578 10926 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0510 15:39:40.152590 10926 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0510 15:39:40.152601 10926 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0510 15:39:40.152612 10926 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0510 15:39:40.152624 10926 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0510 15:39:40.152636 10926 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0510 15:39:40.152647 10926 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.755682 | |
| I0510 15:39:40.152658 10926 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.177778 | |
| I0510 15:39:40.152673 10926 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.72676 (* 0.3 = 1.11803 loss) | |
| I0510 15:39:40.152686 10926 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.1555 (* 0.3 = 0.34665 loss) | |
| I0510 15:39:40.152700 10926 solver.cpp:245] Train net output #76: loss2/loss01 = 3.83381 (* 0.0272727 = 0.104558 loss) | |
| I0510 15:39:40.152717 10926 solver.cpp:245] Train net output #77: loss2/loss02 = 3.80438 (* 0.0272727 = 0.103756 loss) | |
| I0510 15:39:40.152732 10926 solver.cpp:245] Train net output #78: loss2/loss03 = 3.71337 (* 0.0272727 = 0.101274 loss) | |
| I0510 15:39:40.152762 10926 solver.cpp:245] Train net output #79: loss2/loss04 = 4.09657 (* 0.0272727 = 0.111725 loss) | |
| I0510 15:39:40.152778 10926 solver.cpp:245] Train net output #80: loss2/loss05 = 3.27249 (* 0.0272727 = 0.0892498 loss) | |
| I0510 15:39:40.152792 10926 solver.cpp:245] Train net output #81: loss2/loss06 = 2.93769 (* 0.0272727 = 0.0801187 loss) | |
| I0510 15:39:40.152807 10926 solver.cpp:245] Train net output #82: loss2/loss07 = 1.03994 (* 0.0272727 = 0.0283619 loss) | |
| I0510 15:39:40.152822 10926 solver.cpp:245] Train net output #83: loss2/loss08 = 0.690166 (* 0.0272727 = 0.0188227 loss) | |
| I0510 15:39:40.152835 10926 solver.cpp:245] Train net output #84: loss2/loss09 = 0.159766 (* 0.0272727 = 0.00435726 loss) | |
| I0510 15:39:40.152850 10926 solver.cpp:245] Train net output #85: loss2/loss10 = 0.0967383 (* 0.0272727 = 0.00263832 loss) | |
| I0510 15:39:40.152864 10926 solver.cpp:245] Train net output #86: loss2/loss11 = 0.0917948 (* 0.0272727 = 0.00250349 loss) | |
| I0510 15:39:40.152878 10926 solver.cpp:245] Train net output #87: loss2/loss12 = 0.0580612 (* 0.0272727 = 0.00158349 loss) | |
| I0510 15:39:40.152891 10926 solver.cpp:245] Train net output #88: loss2/loss13 = 0.0437339 (* 0.0272727 = 0.00119274 loss) | |
| I0510 15:39:40.152905 10926 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0437965 (* 0.0272727 = 0.00119445 loss) | |
| I0510 15:39:40.152920 10926 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0374391 (* 0.0272727 = 0.00102107 loss) | |
| I0510 15:39:40.152936 10926 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0230391 (* 0.0272727 = 0.000628338 loss) | |
| I0510 15:39:40.152951 10926 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0347764 (* 0.0272727 = 0.000948448 loss) | |
| I0510 15:39:40.152964 10926 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0136744 (* 0.0272727 = 0.000372939 loss) | |
| I0510 15:39:40.152978 10926 solver.cpp:245] Train net output #94: loss2/loss19 = 0.022147 (* 0.0272727 = 0.000604008 loss) | |
| I0510 15:39:40.152993 10926 solver.cpp:245] Train net output #95: loss2/loss20 = 0.0186916 (* 0.0272727 = 0.00050977 loss) | |
| I0510 15:39:40.153007 10926 solver.cpp:245] Train net output #96: loss2/loss21 = 0.0198296 (* 0.0272727 = 0.000540807 loss) | |
| I0510 15:39:40.153020 10926 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0247559 (* 0.0272727 = 0.000675161 loss) | |
| I0510 15:39:40.153033 10926 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0444444 | |
| I0510 15:39:40.153045 10926 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.125 | |
| I0510 15:39:40.153058 10926 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.125 | |
| I0510 15:39:40.153069 10926 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0.125 | |
| I0510 15:39:40.153080 10926 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0510 15:39:40.153092 10926 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.125 | |
| I0510 15:39:40.153103 10926 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.375 | |
| I0510 15:39:40.153115 10926 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.875 | |
| I0510 15:39:40.153146 10926 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.875 | |
| I0510 15:39:40.153158 10926 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0510 15:39:40.153170 10926 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0510 15:39:40.153182 10926 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0510 15:39:40.153193 10926 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0510 15:39:40.153205 10926 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0510 15:39:40.153216 10926 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0510 15:39:40.153228 10926 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0510 15:39:40.153239 10926 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0510 15:39:40.153262 10926 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0510 15:39:40.153275 10926 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0510 15:39:40.153287 10926 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0510 15:39:40.153298 10926 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0510 15:39:40.153309 10926 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0510 15:39:40.153321 10926 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0510 15:39:40.153332 10926 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.744318 | |
| I0510 15:39:40.153343 10926 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.155556 | |
| I0510 15:39:40.153357 10926 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.62564 (* 1 = 3.62564 loss) | |
| I0510 15:39:40.153373 10926 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.22928 (* 1 = 1.22928 loss) | |
| I0510 15:39:40.153383 10926 solver.cpp:245] Train net output #125: loss3/loss01 = 3.49973 (* 0.0909091 = 0.318157 loss) | |
| I0510 15:39:40.153396 10926 solver.cpp:245] Train net output #126: loss3/loss02 = 3.57939 (* 0.0909091 = 0.325399 loss) | |
| I0510 15:39:40.153410 10926 solver.cpp:245] Train net output #127: loss3/loss03 = 3.53315 (* 0.0909091 = 0.321196 loss) | |
| I0510 15:39:40.153424 10926 solver.cpp:245] Train net output #128: loss3/loss04 = 3.7653 (* 0.0909091 = 0.3423 loss) | |
| I0510 15:39:40.153439 10926 solver.cpp:245] Train net output #129: loss3/loss05 = 3.25251 (* 0.0909091 = 0.295683 loss) | |
| I0510 15:39:40.153451 10926 solver.cpp:245] Train net output #130: loss3/loss06 = 2.63713 (* 0.0909091 = 0.239739 loss) | |
| I0510 15:39:40.153465 10926 solver.cpp:245] Train net output #131: loss3/loss07 = 0.960932 (* 0.0909091 = 0.0873575 loss) | |
| I0510 15:39:40.153480 10926 solver.cpp:245] Train net output #132: loss3/loss08 = 0.803031 (* 0.0909091 = 0.0730028 loss) | |
| I0510 15:39:40.153493 10926 solver.cpp:245] Train net output #133: loss3/loss09 = 0.0907082 (* 0.0909091 = 0.0082462 loss) | |
| I0510 15:39:40.153506 10926 solver.cpp:245] Train net output #134: loss3/loss10 = 0.0765699 (* 0.0909091 = 0.0069609 loss) | |
| I0510 15:39:40.153520 10926 solver.cpp:245] Train net output #135: loss3/loss11 = 0.0645931 (* 0.0909091 = 0.0058721 loss) | |
| I0510 15:39:40.153534 10926 solver.cpp:245] Train net output #136: loss3/loss12 = 0.0611856 (* 0.0909091 = 0.00556233 loss) | |
| I0510 15:39:40.153548 10926 solver.cpp:245] Train net output #137: loss3/loss13 = 0.043474 (* 0.0909091 = 0.00395218 loss) | |
| I0510 15:39:40.153563 10926 solver.cpp:245] Train net output #138: loss3/loss14 = 0.0486221 (* 0.0909091 = 0.00442019 loss) | |
| I0510 15:39:40.153576 10926 solver.cpp:245] Train net output #139: loss3/loss15 = 0.023408 (* 0.0909091 = 0.002128 loss) | |
| I0510 15:39:40.153590 10926 solver.cpp:245] Train net output #140: loss3/loss16 = 0.0209752 (* 0.0909091 = 0.00190684 loss) | |
| I0510 15:39:40.153604 10926 solver.cpp:245] Train net output #141: loss3/loss17 = 0.010996 (* 0.0909091 = 0.000999637 loss) | |
| I0510 15:39:40.153619 10926 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00711922 (* 0.0909091 = 0.000647202 loss) | |
| I0510 15:39:40.153631 10926 solver.cpp:245] Train net output #143: loss3/loss19 = 0.0064711 (* 0.0909091 = 0.000588282 loss) | |
| I0510 15:39:40.153645 10926 solver.cpp:245] Train net output #144: loss3/loss20 = 0.0051162 (* 0.0909091 = 0.000465109 loss) | |
| I0510 15:39:40.153659 10926 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00480086 (* 0.0909091 = 0.000436442 loss) | |
| I0510 15:39:40.153673 10926 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00285901 (* 0.0909091 = 0.00025991 loss) | |
| I0510 15:39:40.153686 10926 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0510 15:39:40.153697 10926 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0510 15:39:40.153708 10926 solver.cpp:245] Train net output #149: total_confidence = 9.03318e-10 | |
| I0510 15:39:40.153729 10926 solver.cpp:245] Train net output #150: total_confidence_not_rec = 2.44735e-06 | |
| I0510 15:39:40.153744 10926 sgd_solver.cpp:106] Iteration 3000, lr = 0.001 | |
| I0510 15:42:06.935590 10926 solver.cpp:229] Iteration 3500, loss = 12.5411 | |
| I0510 15:42:06.935724 10926 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0217391 | |
| I0510 15:42:06.935755 10926 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.125 | |
| I0510 15:42:06.935780 10926 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0510 15:42:06.935801 10926 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.25 | |
| I0510 15:42:06.935823 10926 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.125 | |
| I0510 15:42:06.935847 10926 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.25 | |
| I0510 15:42:06.935869 10926 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.375 | |
| I0510 15:42:06.935897 10926 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.625 | |
| I0510 15:42:06.935921 10926 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.875 | |
| I0510 15:42:06.935945 10926 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0510 15:42:06.935966 10926 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0510 15:42:06.935987 10926 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0510 15:42:06.936008 10926 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0510 15:42:06.936029 10926 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0510 15:42:06.936049 10926 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0510 15:42:06.936069 10926 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0510 15:42:06.936091 10926 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0510 15:42:06.936112 10926 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0510 15:42:06.936133 10926 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0510 15:42:06.936154 10926 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0510 15:42:06.936175 10926 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0510 15:42:06.936197 10926 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0510 15:42:06.936216 10926 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0510 15:42:06.936238 10926 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.744318 | |
| I0510 15:42:06.936260 10926 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.130435 | |
| I0510 15:42:06.936288 10926 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 4.09244 (* 0.3 = 1.22773 loss) | |
| I0510 15:42:06.936314 10926 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.23514 (* 0.3 = 0.370542 loss) | |
| I0510 15:42:06.936344 10926 solver.cpp:245] Train net output #27: loss1/loss01 = 3.99875 (* 0.0272727 = 0.109057 loss) | |
| I0510 15:42:06.936375 10926 solver.cpp:245] Train net output #28: loss1/loss02 = 4.38997 (* 0.0272727 = 0.119726 loss) | |
| I0510 15:42:06.936403 10926 solver.cpp:245] Train net output #29: loss1/loss03 = 4.09668 (* 0.0272727 = 0.111728 loss) | |
| I0510 15:42:06.936430 10926 solver.cpp:245] Train net output #30: loss1/loss04 = 4.10541 (* 0.0272727 = 0.111966 loss) | |
| I0510 15:42:06.936455 10926 solver.cpp:245] Train net output #31: loss1/loss05 = 3.23922 (* 0.0272727 = 0.0883425 loss) | |
| I0510 15:42:06.936480 10926 solver.cpp:245] Train net output #32: loss1/loss06 = 3.06052 (* 0.0272727 = 0.0834687 loss) | |
| I0510 15:42:06.936506 10926 solver.cpp:245] Train net output #33: loss1/loss07 = 2.34418 (* 0.0272727 = 0.0639321 loss) | |
| I0510 15:42:06.936530 10926 solver.cpp:245] Train net output #34: loss1/loss08 = 0.883664 (* 0.0272727 = 0.0240999 loss) | |
| I0510 15:42:06.936558 10926 solver.cpp:245] Train net output #35: loss1/loss09 = 0.142058 (* 0.0272727 = 0.00387431 loss) | |
| I0510 15:42:06.936583 10926 solver.cpp:245] Train net output #36: loss1/loss10 = 0.118366 (* 0.0272727 = 0.00322815 loss) | |
| I0510 15:42:06.936609 10926 solver.cpp:245] Train net output #37: loss1/loss11 = 0.057493 (* 0.0272727 = 0.00156799 loss) | |
| I0510 15:42:06.936635 10926 solver.cpp:245] Train net output #38: loss1/loss12 = 0.0638668 (* 0.0272727 = 0.00174182 loss) | |
| I0510 15:42:06.936661 10926 solver.cpp:245] Train net output #39: loss1/loss13 = 0.072828 (* 0.0272727 = 0.00198622 loss) | |
| I0510 15:42:06.936710 10926 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0348138 (* 0.0272727 = 0.000949468 loss) | |
| I0510 15:42:06.936743 10926 solver.cpp:245] Train net output #41: loss1/loss15 = 0.042876 (* 0.0272727 = 0.00116934 loss) | |
| I0510 15:42:06.936771 10926 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0199803 (* 0.0272727 = 0.000544918 loss) | |
| I0510 15:42:06.936797 10926 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0257688 (* 0.0272727 = 0.000702786 loss) | |
| I0510 15:42:06.936823 10926 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0306661 (* 0.0272727 = 0.000836348 loss) | |
| I0510 15:42:06.936849 10926 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0166474 (* 0.0272727 = 0.000454021 loss) | |
| I0510 15:42:06.936875 10926 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0186752 (* 0.0272727 = 0.000509322 loss) | |
| I0510 15:42:06.936902 10926 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0128019 (* 0.0272727 = 0.000349142 loss) | |
| I0510 15:42:06.936930 10926 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0121841 (* 0.0272727 = 0.000332294 loss) | |
| I0510 15:42:06.936954 10926 solver.cpp:245] Train net output #49: loss2/accuracy = 0 | |
| I0510 15:42:06.936976 10926 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0510 15:42:06.936996 10926 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0510 15:42:06.937018 10926 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0510 15:42:06.937039 10926 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.125 | |
| I0510 15:42:06.937062 10926 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.25 | |
| I0510 15:42:06.937084 10926 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.375 | |
| I0510 15:42:06.937105 10926 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.625 | |
| I0510 15:42:06.937149 10926 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.875 | |
| I0510 15:42:06.937175 10926 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0510 15:42:06.937196 10926 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0510 15:42:06.937218 10926 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0510 15:42:06.937239 10926 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0510 15:42:06.937260 10926 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0510 15:42:06.937281 10926 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0510 15:42:06.937302 10926 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0510 15:42:06.937324 10926 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0510 15:42:06.937345 10926 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0510 15:42:06.937366 10926 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0510 15:42:06.937386 10926 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0510 15:42:06.937407 10926 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0510 15:42:06.937429 10926 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0510 15:42:06.937450 10926 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0510 15:42:06.937471 10926 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.727273 | |
| I0510 15:42:06.937494 10926 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.0869565 | |
| I0510 15:42:06.937520 10926 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 4.14687 (* 0.3 = 1.24406 loss) | |
| I0510 15:42:06.937546 10926 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.32661 (* 0.3 = 0.397984 loss) | |
| I0510 15:42:06.937571 10926 solver.cpp:245] Train net output #76: loss2/loss01 = 3.92806 (* 0.0272727 = 0.107129 loss) | |
| I0510 15:42:06.937597 10926 solver.cpp:245] Train net output #77: loss2/loss02 = 3.9536 (* 0.0272727 = 0.107826 loss) | |
| I0510 15:42:06.937643 10926 solver.cpp:245] Train net output #78: loss2/loss03 = 4.20011 (* 0.0272727 = 0.114548 loss) | |
| I0510 15:42:06.937669 10926 solver.cpp:245] Train net output #79: loss2/loss04 = 4.23765 (* 0.0272727 = 0.115572 loss) | |
| I0510 15:42:06.937695 10926 solver.cpp:245] Train net output #80: loss2/loss05 = 3.64902 (* 0.0272727 = 0.0995189 loss) | |
| I0510 15:42:06.937721 10926 solver.cpp:245] Train net output #81: loss2/loss06 = 3.06644 (* 0.0272727 = 0.0836302 loss) | |
| I0510 15:42:06.937747 10926 solver.cpp:245] Train net output #82: loss2/loss07 = 1.75692 (* 0.0272727 = 0.0479161 loss) | |
| I0510 15:42:06.937772 10926 solver.cpp:245] Train net output #83: loss2/loss08 = 0.852372 (* 0.0272727 = 0.0232465 loss) | |
| I0510 15:42:06.937803 10926 solver.cpp:245] Train net output #84: loss2/loss09 = 0.13253 (* 0.0272727 = 0.00361444 loss) | |
| I0510 15:42:06.937832 10926 solver.cpp:245] Train net output #85: loss2/loss10 = 0.0893677 (* 0.0272727 = 0.0024373 loss) | |
| I0510 15:42:06.937865 10926 solver.cpp:245] Train net output #86: loss2/loss11 = 0.0891238 (* 0.0272727 = 0.00243065 loss) | |
| I0510 15:42:06.937892 10926 solver.cpp:245] Train net output #87: loss2/loss12 = 0.0468821 (* 0.0272727 = 0.0012786 loss) | |
| I0510 15:42:06.937919 10926 solver.cpp:245] Train net output #88: loss2/loss13 = 0.0487679 (* 0.0272727 = 0.00133003 loss) | |
| I0510 15:42:06.937945 10926 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0288709 (* 0.0272727 = 0.000787389 loss) | |
| I0510 15:42:06.937973 10926 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0272018 (* 0.0272727 = 0.000741868 loss) | |
| I0510 15:42:06.938002 10926 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0275215 (* 0.0272727 = 0.000750587 loss) | |
| I0510 15:42:06.938030 10926 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0180214 (* 0.0272727 = 0.000491494 loss) | |
| I0510 15:42:06.938055 10926 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0116921 (* 0.0272727 = 0.000318875 loss) | |
| I0510 15:42:06.938081 10926 solver.cpp:245] Train net output #94: loss2/loss19 = 0.0120899 (* 0.0272727 = 0.000329726 loss) | |
| I0510 15:42:06.938107 10926 solver.cpp:245] Train net output #95: loss2/loss20 = 0.0166784 (* 0.0272727 = 0.000454866 loss) | |
| I0510 15:42:06.938133 10926 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00906773 (* 0.0272727 = 0.000247302 loss) | |
| I0510 15:42:06.938158 10926 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0159067 (* 0.0272727 = 0.000433819 loss) | |
| I0510 15:42:06.938180 10926 solver.cpp:245] Train net output #98: loss3/accuracy = 0 | |
| I0510 15:42:06.938201 10926 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0510 15:42:06.938222 10926 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0510 15:42:06.938242 10926 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0510 15:42:06.938264 10926 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0510 15:42:06.938287 10926 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.25 | |
| I0510 15:42:06.938307 10926 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.375 | |
| I0510 15:42:06.938326 10926 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.625 | |
| I0510 15:42:06.938347 10926 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.875 | |
| I0510 15:42:06.938369 10926 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0510 15:42:06.938390 10926 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0510 15:42:06.938410 10926 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0510 15:42:06.938431 10926 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0510 15:42:06.938452 10926 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0510 15:42:06.938472 10926 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0510 15:42:06.938493 10926 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0510 15:42:06.938513 10926 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0510 15:42:06.938550 10926 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0510 15:42:06.938573 10926 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0510 15:42:06.938594 10926 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0510 15:42:06.938616 10926 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0510 15:42:06.938637 10926 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0510 15:42:06.938657 10926 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0510 15:42:06.938678 10926 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.732955 | |
| I0510 15:42:06.938699 10926 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.130435 | |
| I0510 15:42:06.938724 10926 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.93352 (* 1 = 3.93352 loss) | |
| I0510 15:42:06.938750 10926 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.19313 (* 1 = 1.19313 loss) | |
| I0510 15:42:06.938774 10926 solver.cpp:245] Train net output #125: loss3/loss01 = 3.8755 (* 0.0909091 = 0.352318 loss) | |
| I0510 15:42:06.938799 10926 solver.cpp:245] Train net output #126: loss3/loss02 = 4.06916 (* 0.0909091 = 0.369924 loss) | |
| I0510 15:42:06.938827 10926 solver.cpp:245] Train net output #127: loss3/loss03 = 3.95901 (* 0.0909091 = 0.35991 loss) | |
| I0510 15:42:06.938855 10926 solver.cpp:245] Train net output #128: loss3/loss04 = 3.78376 (* 0.0909091 = 0.343979 loss) | |
| I0510 15:42:06.938881 10926 solver.cpp:245] Train net output #129: loss3/loss05 = 2.94614 (* 0.0909091 = 0.267831 loss) | |
| I0510 15:42:06.938907 10926 solver.cpp:245] Train net output #130: loss3/loss06 = 3.09552 (* 0.0909091 = 0.28141 loss) | |
| I0510 15:42:06.938932 10926 solver.cpp:245] Train net output #131: loss3/loss07 = 1.96508 (* 0.0909091 = 0.178643 loss) | |
| I0510 15:42:06.938957 10926 solver.cpp:245] Train net output #132: loss3/loss08 = 1.02645 (* 0.0909091 = 0.0933137 loss) | |
| I0510 15:42:06.938984 10926 solver.cpp:245] Train net output #133: loss3/loss09 = 0.132967 (* 0.0909091 = 0.0120879 loss) | |
| I0510 15:42:06.939010 10926 solver.cpp:245] Train net output #134: loss3/loss10 = 0.0982752 (* 0.0909091 = 0.00893411 loss) | |
| I0510 15:42:06.939038 10926 solver.cpp:245] Train net output #135: loss3/loss11 = 0.0882422 (* 0.0909091 = 0.00802202 loss) | |
| I0510 15:42:06.939066 10926 solver.cpp:245] Train net output #136: loss3/loss12 = 0.0708675 (* 0.0909091 = 0.0064425 loss) | |
| I0510 15:42:06.939092 10926 solver.cpp:245] Train net output #137: loss3/loss13 = 0.0513587 (* 0.0909091 = 0.00466897 loss) | |
| I0510 15:42:06.939117 10926 solver.cpp:245] Train net output #138: loss3/loss14 = 0.0440804 (* 0.0909091 = 0.00400731 loss) | |
| I0510 15:42:06.939143 10926 solver.cpp:245] Train net output #139: loss3/loss15 = 0.0310656 (* 0.0909091 = 0.00282415 loss) | |
| I0510 15:42:06.939168 10926 solver.cpp:245] Train net output #140: loss3/loss16 = 0.0179491 (* 0.0909091 = 0.00163173 loss) | |
| I0510 15:42:06.939193 10926 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00597154 (* 0.0909091 = 0.000542867 loss) | |
| I0510 15:42:06.939219 10926 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00649312 (* 0.0909091 = 0.000590284 loss) | |
| I0510 15:42:06.939244 10926 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00406844 (* 0.0909091 = 0.000369858 loss) | |
| I0510 15:42:06.939270 10926 solver.cpp:245] Train net output #144: loss3/loss20 = 0.0034813 (* 0.0909091 = 0.000316482 loss) | |
| I0510 15:42:06.939296 10926 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00347777 (* 0.0909091 = 0.000316161 loss) | |
| I0510 15:42:06.939322 10926 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00198751 (* 0.0909091 = 0.000180683 loss) | |
| I0510 15:42:06.939344 10926 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0510 15:42:06.939365 10926 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0510 15:42:06.939386 10926 solver.cpp:245] Train net output #149: total_confidence = 4.68174e-08 | |
| I0510 15:42:06.939424 10926 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.000130565 | |
| I0510 15:42:06.939447 10926 sgd_solver.cpp:106] Iteration 3500, lr = 0.001 | |
| I0510 15:44:33.875632 10926 solver.cpp:229] Iteration 4000, loss = 12.3095 | |
| I0510 15:44:33.875774 10926 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0789474 | |
| I0510 15:44:33.875794 10926 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0510 15:44:33.875808 10926 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.25 | |
| I0510 15:44:33.875820 10926 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0510 15:44:33.875833 10926 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.125 | |
| I0510 15:44:33.875844 10926 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.25 | |
| I0510 15:44:33.875856 10926 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.75 | |
| I0510 15:44:33.875869 10926 solver.cpp:245] Train net output #7: loss1/accuracy07 = 1 | |
| I0510 15:44:33.875882 10926 solver.cpp:245] Train net output #8: loss1/accuracy08 = 1 | |
| I0510 15:44:33.875895 10926 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0510 15:44:33.875906 10926 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0510 15:44:33.875917 10926 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0510 15:44:33.875929 10926 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0510 15:44:33.875941 10926 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0510 15:44:33.875952 10926 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0510 15:44:33.875963 10926 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0510 15:44:33.875975 10926 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0510 15:44:33.875988 10926 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0510 15:44:33.875999 10926 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0510 15:44:33.876010 10926 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0510 15:44:33.876021 10926 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0510 15:44:33.876034 10926 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0510 15:44:33.876044 10926 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0510 15:44:33.876056 10926 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.795455 | |
| I0510 15:44:33.876068 10926 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.131579 | |
| I0510 15:44:33.876085 10926 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.87768 (* 0.3 = 1.1633 loss) | |
| I0510 15:44:33.876098 10926 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.06601 (* 0.3 = 0.319804 loss) | |
| I0510 15:44:33.876121 10926 solver.cpp:245] Train net output #27: loss1/loss01 = 4.22742 (* 0.0272727 = 0.115293 loss) | |
| I0510 15:44:33.876137 10926 solver.cpp:245] Train net output #28: loss1/loss02 = 3.43744 (* 0.0272727 = 0.0937483 loss) | |
| I0510 15:44:33.876150 10926 solver.cpp:245] Train net output #29: loss1/loss03 = 4.23741 (* 0.0272727 = 0.115566 loss) | |
| I0510 15:44:33.876164 10926 solver.cpp:245] Train net output #30: loss1/loss04 = 3.17445 (* 0.0272727 = 0.086576 loss) | |
| I0510 15:44:33.876178 10926 solver.cpp:245] Train net output #31: loss1/loss05 = 3.88277 (* 0.0272727 = 0.105894 loss) | |
| I0510 15:44:33.876193 10926 solver.cpp:245] Train net output #32: loss1/loss06 = 1.65651 (* 0.0272727 = 0.0451775 loss) | |
| I0510 15:44:33.876206 10926 solver.cpp:245] Train net output #33: loss1/loss07 = 0.376481 (* 0.0272727 = 0.0102677 loss) | |
| I0510 15:44:33.876220 10926 solver.cpp:245] Train net output #34: loss1/loss08 = 0.198915 (* 0.0272727 = 0.00542496 loss) | |
| I0510 15:44:33.876235 10926 solver.cpp:245] Train net output #35: loss1/loss09 = 0.167311 (* 0.0272727 = 0.00456302 loss) | |
| I0510 15:44:33.876250 10926 solver.cpp:245] Train net output #36: loss1/loss10 = 0.121488 (* 0.0272727 = 0.00331332 loss) | |
| I0510 15:44:33.876263 10926 solver.cpp:245] Train net output #37: loss1/loss11 = 0.088237 (* 0.0272727 = 0.00240646 loss) | |
| I0510 15:44:33.876277 10926 solver.cpp:245] Train net output #38: loss1/loss12 = 0.0555589 (* 0.0272727 = 0.00151524 loss) | |
| I0510 15:44:33.876291 10926 solver.cpp:245] Train net output #39: loss1/loss13 = 0.0902547 (* 0.0272727 = 0.00246149 loss) | |
| I0510 15:44:33.876323 10926 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0567824 (* 0.0272727 = 0.00154861 loss) | |
| I0510 15:44:33.876339 10926 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0345627 (* 0.0272727 = 0.000942618 loss) | |
| I0510 15:44:33.876354 10926 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0233588 (* 0.0272727 = 0.000637059 loss) | |
| I0510 15:44:33.876368 10926 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0238629 (* 0.0272727 = 0.000650807 loss) | |
| I0510 15:44:33.876382 10926 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0219253 (* 0.0272727 = 0.000597962 loss) | |
| I0510 15:44:33.876396 10926 solver.cpp:245] Train net output #45: loss1/loss19 = 0.020749 (* 0.0272727 = 0.000565881 loss) | |
| I0510 15:44:33.876410 10926 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0190977 (* 0.0272727 = 0.000520847 loss) | |
| I0510 15:44:33.876425 10926 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0159781 (* 0.0272727 = 0.000435766 loss) | |
| I0510 15:44:33.876438 10926 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0240673 (* 0.0272727 = 0.00065638 loss) | |
| I0510 15:44:33.876451 10926 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0526316 | |
| I0510 15:44:33.876462 10926 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0510 15:44:33.876474 10926 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0.25 | |
| I0510 15:44:33.876487 10926 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0.125 | |
| I0510 15:44:33.876498 10926 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.375 | |
| I0510 15:44:33.876509 10926 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.25 | |
| I0510 15:44:33.876520 10926 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.75 | |
| I0510 15:44:33.876533 10926 solver.cpp:245] Train net output #56: loss2/accuracy07 = 1 | |
| I0510 15:44:33.876543 10926 solver.cpp:245] Train net output #57: loss2/accuracy08 = 1 | |
| I0510 15:44:33.876554 10926 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0510 15:44:33.876566 10926 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0510 15:44:33.876577 10926 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0510 15:44:33.876588 10926 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0510 15:44:33.876600 10926 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0510 15:44:33.876611 10926 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0510 15:44:33.876622 10926 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0510 15:44:33.876634 10926 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0510 15:44:33.876646 10926 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0510 15:44:33.876657 10926 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0510 15:44:33.876664 10926 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0510 15:44:33.876672 10926 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0510 15:44:33.876684 10926 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0510 15:44:33.876696 10926 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0510 15:44:33.876708 10926 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.789773 | |
| I0510 15:44:33.876719 10926 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.105263 | |
| I0510 15:44:33.876734 10926 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.80225 (* 0.3 = 1.14068 loss) | |
| I0510 15:44:33.876754 10926 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.13253 (* 0.3 = 0.33976 loss) | |
| I0510 15:44:33.876770 10926 solver.cpp:245] Train net output #76: loss2/loss01 = 3.61905 (* 0.0272727 = 0.0987013 loss) | |
| I0510 15:44:33.876787 10926 solver.cpp:245] Train net output #77: loss2/loss02 = 3.43578 (* 0.0272727 = 0.0937031 loss) | |
| I0510 15:44:33.876802 10926 solver.cpp:245] Train net output #78: loss2/loss03 = 3.47409 (* 0.0272727 = 0.0947479 loss) | |
| I0510 15:44:33.876828 10926 solver.cpp:245] Train net output #79: loss2/loss04 = 3.56219 (* 0.0272727 = 0.0971506 loss) | |
| I0510 15:44:33.876843 10926 solver.cpp:245] Train net output #80: loss2/loss05 = 3.5267 (* 0.0272727 = 0.0961827 loss) | |
| I0510 15:44:33.876857 10926 solver.cpp:245] Train net output #81: loss2/loss06 = 1.72688 (* 0.0272727 = 0.0470968 loss) | |
| I0510 15:44:33.876871 10926 solver.cpp:245] Train net output #82: loss2/loss07 = 0.645332 (* 0.0272727 = 0.0176 loss) | |
| I0510 15:44:33.876885 10926 solver.cpp:245] Train net output #83: loss2/loss08 = 0.360586 (* 0.0272727 = 0.00983416 loss) | |
| I0510 15:44:33.876899 10926 solver.cpp:245] Train net output #84: loss2/loss09 = 0.133072 (* 0.0272727 = 0.00362923 loss) | |
| I0510 15:44:33.876914 10926 solver.cpp:245] Train net output #85: loss2/loss10 = 0.175151 (* 0.0272727 = 0.00477685 loss) | |
| I0510 15:44:33.876929 10926 solver.cpp:245] Train net output #86: loss2/loss11 = 0.142503 (* 0.0272727 = 0.00388644 loss) | |
| I0510 15:44:33.876945 10926 solver.cpp:245] Train net output #87: loss2/loss12 = 0.121572 (* 0.0272727 = 0.0033156 loss) | |
| I0510 15:44:33.876958 10926 solver.cpp:245] Train net output #88: loss2/loss13 = 0.0854207 (* 0.0272727 = 0.00232966 loss) | |
| I0510 15:44:33.876972 10926 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0687286 (* 0.0272727 = 0.00187442 loss) | |
| I0510 15:44:33.876986 10926 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0918032 (* 0.0272727 = 0.00250372 loss) | |
| I0510 15:44:33.877001 10926 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0578789 (* 0.0272727 = 0.00157851 loss) | |
| I0510 15:44:33.877014 10926 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0353432 (* 0.0272727 = 0.000963905 loss) | |
| I0510 15:44:33.877028 10926 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0405102 (* 0.0272727 = 0.00110482 loss) | |
| I0510 15:44:33.877043 10926 solver.cpp:245] Train net output #94: loss2/loss19 = 0.0279695 (* 0.0272727 = 0.000762804 loss) | |
| I0510 15:44:33.877055 10926 solver.cpp:245] Train net output #95: loss2/loss20 = 0.0465366 (* 0.0272727 = 0.00126918 loss) | |
| I0510 15:44:33.877069 10926 solver.cpp:245] Train net output #96: loss2/loss21 = 0.0265894 (* 0.0272727 = 0.000725165 loss) | |
| I0510 15:44:33.877084 10926 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0231018 (* 0.0272727 = 0.000630048 loss) | |
| I0510 15:44:33.877095 10926 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0526316 | |
| I0510 15:44:33.877106 10926 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.125 | |
| I0510 15:44:33.877130 10926 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0510 15:44:33.877145 10926 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0510 15:44:33.877156 10926 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.25 | |
| I0510 15:44:33.877167 10926 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.25 | |
| I0510 15:44:33.877179 10926 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.75 | |
| I0510 15:44:33.877190 10926 solver.cpp:245] Train net output #105: loss3/accuracy07 = 1 | |
| I0510 15:44:33.877202 10926 solver.cpp:245] Train net output #106: loss3/accuracy08 = 1 | |
| I0510 15:44:33.877213 10926 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0510 15:44:33.877225 10926 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0510 15:44:33.877236 10926 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0510 15:44:33.877248 10926 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0510 15:44:33.877259 10926 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0510 15:44:33.877270 10926 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0510 15:44:33.877281 10926 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0510 15:44:33.877292 10926 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0510 15:44:33.877316 10926 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0510 15:44:33.877329 10926 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0510 15:44:33.877341 10926 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0510 15:44:33.877352 10926 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0510 15:44:33.877363 10926 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0510 15:44:33.877374 10926 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0510 15:44:33.877387 10926 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.772727 | |
| I0510 15:44:33.877398 10926 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.157895 | |
| I0510 15:44:33.877411 10926 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.55321 (* 1 = 3.55321 loss) | |
| I0510 15:44:33.877425 10926 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.07749 (* 1 = 1.07749 loss) | |
| I0510 15:44:33.877439 10926 solver.cpp:245] Train net output #125: loss3/loss01 = 3.58334 (* 0.0909091 = 0.325758 loss) | |
| I0510 15:44:33.877454 10926 solver.cpp:245] Train net output #126: loss3/loss02 = 3.37137 (* 0.0909091 = 0.306489 loss) | |
| I0510 15:44:33.877466 10926 solver.cpp:245] Train net output #127: loss3/loss03 = 3.46266 (* 0.0909091 = 0.314787 loss) | |
| I0510 15:44:33.877480 10926 solver.cpp:245] Train net output #128: loss3/loss04 = 3.0308 (* 0.0909091 = 0.275527 loss) | |
| I0510 15:44:33.877495 10926 solver.cpp:245] Train net output #129: loss3/loss05 = 3.24876 (* 0.0909091 = 0.295342 loss) | |
| I0510 15:44:33.877507 10926 solver.cpp:245] Train net output #130: loss3/loss06 = 1.56081 (* 0.0909091 = 0.141892 loss) | |
| I0510 15:44:33.877521 10926 solver.cpp:245] Train net output #131: loss3/loss07 = 0.389262 (* 0.0909091 = 0.0353875 loss) | |
| I0510 15:44:33.877535 10926 solver.cpp:245] Train net output #132: loss3/loss08 = 0.207919 (* 0.0909091 = 0.0189018 loss) | |
| I0510 15:44:33.877549 10926 solver.cpp:245] Train net output #133: loss3/loss09 = 0.105583 (* 0.0909091 = 0.0095985 loss) | |
| I0510 15:44:33.877564 10926 solver.cpp:245] Train net output #134: loss3/loss10 = 0.0816778 (* 0.0909091 = 0.00742525 loss) | |
| I0510 15:44:33.877578 10926 solver.cpp:245] Train net output #135: loss3/loss11 = 0.0726766 (* 0.0909091 = 0.00660696 loss) | |
| I0510 15:44:33.877593 10926 solver.cpp:245] Train net output #136: loss3/loss12 = 0.0406615 (* 0.0909091 = 0.0036965 loss) | |
| I0510 15:44:33.877606 10926 solver.cpp:245] Train net output #137: loss3/loss13 = 0.0354635 (* 0.0909091 = 0.00322395 loss) | |
| I0510 15:44:33.877620 10926 solver.cpp:245] Train net output #138: loss3/loss14 = 0.0338066 (* 0.0909091 = 0.00307333 loss) | |
| I0510 15:44:33.877635 10926 solver.cpp:245] Train net output #139: loss3/loss15 = 0.0262534 (* 0.0909091 = 0.00238667 loss) | |
| I0510 15:44:33.877648 10926 solver.cpp:245] Train net output #140: loss3/loss16 = 0.0151048 (* 0.0909091 = 0.00137316 loss) | |
| I0510 15:44:33.877662 10926 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00835618 (* 0.0909091 = 0.000759653 loss) | |
| I0510 15:44:33.877676 10926 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00882588 (* 0.0909091 = 0.000802352 loss) | |
| I0510 15:44:33.877691 10926 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00695377 (* 0.0909091 = 0.000632161 loss) | |
| I0510 15:44:33.877704 10926 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00593357 (* 0.0909091 = 0.000539415 loss) | |
| I0510 15:44:33.877718 10926 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00400756 (* 0.0909091 = 0.000364324 loss) | |
| I0510 15:44:33.877732 10926 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00402761 (* 0.0909091 = 0.000366147 loss) | |
| I0510 15:44:33.877744 10926 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0510 15:44:33.877755 10926 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0510 15:44:33.877768 10926 solver.cpp:245] Train net output #149: total_confidence = 1.12785e-07 | |
| I0510 15:44:33.877789 10926 solver.cpp:245] Train net output #150: total_confidence_not_rec = 6.90498e-05 | |
| I0510 15:44:33.877804 10926 sgd_solver.cpp:106] Iteration 4000, lr = 0.001 | |
| I0510 15:46:37.616171 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 34.4185 > 30) by scale factor 0.871624 | |
| I0510 15:47:00.747972 10926 solver.cpp:229] Iteration 4500, loss = 12.0919 | |
| I0510 15:47:00.748051 10926 solver.cpp:245] Train net output #0: loss1/accuracy = 0.02 | |
| I0510 15:47:00.748070 10926 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.125 | |
| I0510 15:47:00.748085 10926 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.125 | |
| I0510 15:47:00.748097 10926 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.125 | |
| I0510 15:47:00.748109 10926 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0 | |
| I0510 15:47:00.748122 10926 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.25 | |
| I0510 15:47:00.748134 10926 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.5 | |
| I0510 15:47:00.748147 10926 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.5 | |
| I0510 15:47:00.748159 10926 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.875 | |
| I0510 15:47:00.748172 10926 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.875 | |
| I0510 15:47:00.748184 10926 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.875 | |
| I0510 15:47:00.748196 10926 solver.cpp:245] Train net output #11: loss1/accuracy11 = 0.875 | |
| I0510 15:47:00.748209 10926 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0510 15:47:00.748221 10926 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0510 15:47:00.748234 10926 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0510 15:47:00.748245 10926 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0510 15:47:00.748257 10926 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0510 15:47:00.748270 10926 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0510 15:47:00.748281 10926 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0510 15:47:00.748293 10926 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0510 15:47:00.748306 10926 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0510 15:47:00.748317 10926 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0510 15:47:00.748329 10926 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0510 15:47:00.748342 10926 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.721591 | |
| I0510 15:47:00.748354 10926 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.14 | |
| I0510 15:47:00.748371 10926 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.60321 (* 0.3 = 1.08096 loss) | |
| I0510 15:47:00.748386 10926 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.1893 (* 0.3 = 0.35679 loss) | |
| I0510 15:47:00.748401 10926 solver.cpp:245] Train net output #27: loss1/loss01 = 3.54838 (* 0.0272727 = 0.0967739 loss) | |
| I0510 15:47:00.748415 10926 solver.cpp:245] Train net output #28: loss1/loss02 = 3.35541 (* 0.0272727 = 0.0915111 loss) | |
| I0510 15:47:00.748430 10926 solver.cpp:245] Train net output #29: loss1/loss03 = 4.03958 (* 0.0272727 = 0.11017 loss) | |
| I0510 15:47:00.748445 10926 solver.cpp:245] Train net output #30: loss1/loss04 = 3.76202 (* 0.0272727 = 0.102601 loss) | |
| I0510 15:47:00.748458 10926 solver.cpp:245] Train net output #31: loss1/loss05 = 3.75409 (* 0.0272727 = 0.102384 loss) | |
| I0510 15:47:00.748472 10926 solver.cpp:245] Train net output #32: loss1/loss06 = 2.69471 (* 0.0272727 = 0.0734922 loss) | |
| I0510 15:47:00.748486 10926 solver.cpp:245] Train net output #33: loss1/loss07 = 3.04094 (* 0.0272727 = 0.0829346 loss) | |
| I0510 15:47:00.748502 10926 solver.cpp:245] Train net output #34: loss1/loss08 = 0.751728 (* 0.0272727 = 0.0205017 loss) | |
| I0510 15:47:00.748515 10926 solver.cpp:245] Train net output #35: loss1/loss09 = 0.87556 (* 0.0272727 = 0.0238789 loss) | |
| I0510 15:47:00.748529 10926 solver.cpp:245] Train net output #36: loss1/loss10 = 0.81383 (* 0.0272727 = 0.0221954 loss) | |
| I0510 15:47:00.748543 10926 solver.cpp:245] Train net output #37: loss1/loss11 = 0.788891 (* 0.0272727 = 0.0215152 loss) | |
| I0510 15:47:00.748558 10926 solver.cpp:245] Train net output #38: loss1/loss12 = 0.0850608 (* 0.0272727 = 0.00231984 loss) | |
| I0510 15:47:00.748615 10926 solver.cpp:245] Train net output #39: loss1/loss13 = 0.081562 (* 0.0272727 = 0.00222442 loss) | |
| I0510 15:47:00.748634 10926 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0505575 (* 0.0272727 = 0.00137884 loss) | |
| I0510 15:47:00.748651 10926 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0459445 (* 0.0272727 = 0.00125303 loss) | |
| I0510 15:47:00.748664 10926 solver.cpp:245] Train net output #42: loss1/loss16 = 0.052806 (* 0.0272727 = 0.00144016 loss) | |
| I0510 15:47:00.748679 10926 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0220885 (* 0.0272727 = 0.000602415 loss) | |
| I0510 15:47:00.748693 10926 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0190281 (* 0.0272727 = 0.000518948 loss) | |
| I0510 15:47:00.748708 10926 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0216943 (* 0.0272727 = 0.000591664 loss) | |
| I0510 15:47:00.748723 10926 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0238477 (* 0.0272727 = 0.000650391 loss) | |
| I0510 15:47:00.748739 10926 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0141754 (* 0.0272727 = 0.000386601 loss) | |
| I0510 15:47:00.748754 10926 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0161029 (* 0.0272727 = 0.000439169 loss) | |
| I0510 15:47:00.748767 10926 solver.cpp:245] Train net output #49: loss2/accuracy = 0.08 | |
| I0510 15:47:00.748780 10926 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.25 | |
| I0510 15:47:00.748792 10926 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0.125 | |
| I0510 15:47:00.748805 10926 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0510 15:47:00.748816 10926 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0 | |
| I0510 15:47:00.748827 10926 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.25 | |
| I0510 15:47:00.748839 10926 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.5 | |
| I0510 15:47:00.748852 10926 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.5 | |
| I0510 15:47:00.748863 10926 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.875 | |
| I0510 15:47:00.748875 10926 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.875 | |
| I0510 15:47:00.748888 10926 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.875 | |
| I0510 15:47:00.748899 10926 solver.cpp:245] Train net output #60: loss2/accuracy11 = 0.875 | |
| I0510 15:47:00.748911 10926 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0510 15:47:00.748922 10926 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0510 15:47:00.748934 10926 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0510 15:47:00.748946 10926 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0510 15:47:00.748958 10926 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0510 15:47:00.748970 10926 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0510 15:47:00.748981 10926 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0510 15:47:00.748992 10926 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0510 15:47:00.749004 10926 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0510 15:47:00.749016 10926 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0510 15:47:00.749027 10926 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0510 15:47:00.749039 10926 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.727273 | |
| I0510 15:47:00.749053 10926 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.2 | |
| I0510 15:47:00.749063 10926 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.57498 (* 0.3 = 1.0725 loss) | |
| I0510 15:47:00.749073 10926 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.24817 (* 0.3 = 0.374451 loss) | |
| I0510 15:47:00.749088 10926 solver.cpp:245] Train net output #76: loss2/loss01 = 3.95085 (* 0.0272727 = 0.10775 loss) | |
| I0510 15:47:00.749101 10926 solver.cpp:245] Train net output #77: loss2/loss02 = 3.39161 (* 0.0272727 = 0.0924985 loss) | |
| I0510 15:47:00.749142 10926 solver.cpp:245] Train net output #78: loss2/loss03 = 4.04038 (* 0.0272727 = 0.110192 loss) | |
| I0510 15:47:00.749158 10926 solver.cpp:245] Train net output #79: loss2/loss04 = 3.51288 (* 0.0272727 = 0.0958058 loss) | |
| I0510 15:47:00.749173 10926 solver.cpp:245] Train net output #80: loss2/loss05 = 3.08699 (* 0.0272727 = 0.0841906 loss) | |
| I0510 15:47:00.749187 10926 solver.cpp:245] Train net output #81: loss2/loss06 = 2.28018 (* 0.0272727 = 0.0621867 loss) | |
| I0510 15:47:00.749202 10926 solver.cpp:245] Train net output #82: loss2/loss07 = 2.63502 (* 0.0272727 = 0.0718642 loss) | |
| I0510 15:47:00.749215 10926 solver.cpp:245] Train net output #83: loss2/loss08 = 0.810604 (* 0.0272727 = 0.0221074 loss) | |
| I0510 15:47:00.749230 10926 solver.cpp:245] Train net output #84: loss2/loss09 = 1.08768 (* 0.0272727 = 0.0296639 loss) | |
| I0510 15:47:00.749243 10926 solver.cpp:245] Train net output #85: loss2/loss10 = 0.870296 (* 0.0272727 = 0.0237354 loss) | |
| I0510 15:47:00.749258 10926 solver.cpp:245] Train net output #86: loss2/loss11 = 0.902061 (* 0.0272727 = 0.0246017 loss) | |
| I0510 15:47:00.749272 10926 solver.cpp:245] Train net output #87: loss2/loss12 = 0.0596257 (* 0.0272727 = 0.00162616 loss) | |
| I0510 15:47:00.749286 10926 solver.cpp:245] Train net output #88: loss2/loss13 = 0.0377361 (* 0.0272727 = 0.00102917 loss) | |
| I0510 15:47:00.749301 10926 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0220206 (* 0.0272727 = 0.000600563 loss) | |
| I0510 15:47:00.749316 10926 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0318739 (* 0.0272727 = 0.00086929 loss) | |
| I0510 15:47:00.749330 10926 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0163993 (* 0.0272727 = 0.000447254 loss) | |
| I0510 15:47:00.749346 10926 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0122165 (* 0.0272727 = 0.000333178 loss) | |
| I0510 15:47:00.749359 10926 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0196548 (* 0.0272727 = 0.00053604 loss) | |
| I0510 15:47:00.749374 10926 solver.cpp:245] Train net output #94: loss2/loss19 = 0.0100854 (* 0.0272727 = 0.000275056 loss) | |
| I0510 15:47:00.749388 10926 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00872187 (* 0.0272727 = 0.000237869 loss) | |
| I0510 15:47:00.749402 10926 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00765373 (* 0.0272727 = 0.000208738 loss) | |
| I0510 15:47:00.749416 10926 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0100511 (* 0.0272727 = 0.000274122 loss) | |
| I0510 15:47:00.749429 10926 solver.cpp:245] Train net output #98: loss3/accuracy = 0.12 | |
| I0510 15:47:00.749442 10926 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0510 15:47:00.749454 10926 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0510 15:47:00.749466 10926 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0.125 | |
| I0510 15:47:00.749477 10926 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0510 15:47:00.749490 10926 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.125 | |
| I0510 15:47:00.749501 10926 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.5 | |
| I0510 15:47:00.749513 10926 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.5 | |
| I0510 15:47:00.749526 10926 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.875 | |
| I0510 15:47:00.749537 10926 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.875 | |
| I0510 15:47:00.749549 10926 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.875 | |
| I0510 15:47:00.749560 10926 solver.cpp:245] Train net output #109: loss3/accuracy11 = 0.875 | |
| I0510 15:47:00.749573 10926 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0510 15:47:00.749584 10926 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0510 15:47:00.749596 10926 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0510 15:47:00.749608 10926 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0510 15:47:00.749632 10926 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0510 15:47:00.749645 10926 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0510 15:47:00.749657 10926 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0510 15:47:00.749670 10926 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0510 15:47:00.749683 10926 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0510 15:47:00.749696 10926 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0510 15:47:00.749708 10926 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0510 15:47:00.749721 10926 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.732955 | |
| I0510 15:47:00.749733 10926 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.26 | |
| I0510 15:47:00.749747 10926 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.41048 (* 1 = 3.41048 loss) | |
| I0510 15:47:00.749761 10926 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.30411 (* 1 = 1.30411 loss) | |
| I0510 15:47:00.749775 10926 solver.cpp:245] Train net output #125: loss3/loss01 = 3.87327 (* 0.0909091 = 0.352116 loss) | |
| I0510 15:47:00.749793 10926 solver.cpp:245] Train net output #126: loss3/loss02 = 2.98315 (* 0.0909091 = 0.271195 loss) | |
| I0510 15:47:00.749807 10926 solver.cpp:245] Train net output #127: loss3/loss03 = 3.33738 (* 0.0909091 = 0.303398 loss) | |
| I0510 15:47:00.749821 10926 solver.cpp:245] Train net output #128: loss3/loss04 = 3.15402 (* 0.0909091 = 0.286729 loss) | |
| I0510 15:47:00.749836 10926 solver.cpp:245] Train net output #129: loss3/loss05 = 3.53498 (* 0.0909091 = 0.321362 loss) | |
| I0510 15:47:00.749850 10926 solver.cpp:245] Train net output #130: loss3/loss06 = 2.44545 (* 0.0909091 = 0.222314 loss) | |
| I0510 15:47:00.749864 10926 solver.cpp:245] Train net output #131: loss3/loss07 = 2.65231 (* 0.0909091 = 0.241119 loss) | |
| I0510 15:47:00.749878 10926 solver.cpp:245] Train net output #132: loss3/loss08 = 0.881012 (* 0.0909091 = 0.080092 loss) | |
| I0510 15:47:00.749893 10926 solver.cpp:245] Train net output #133: loss3/loss09 = 1.03906 (* 0.0909091 = 0.0944604 loss) | |
| I0510 15:47:00.749907 10926 solver.cpp:245] Train net output #134: loss3/loss10 = 0.573618 (* 0.0909091 = 0.0521471 loss) | |
| I0510 15:47:00.749922 10926 solver.cpp:245] Train net output #135: loss3/loss11 = 0.6623 (* 0.0909091 = 0.060209 loss) | |
| I0510 15:47:00.749935 10926 solver.cpp:245] Train net output #136: loss3/loss12 = 0.174082 (* 0.0909091 = 0.0158257 loss) | |
| I0510 15:47:00.749950 10926 solver.cpp:245] Train net output #137: loss3/loss13 = 0.100065 (* 0.0909091 = 0.00909679 loss) | |
| I0510 15:47:00.749965 10926 solver.cpp:245] Train net output #138: loss3/loss14 = 0.0538934 (* 0.0909091 = 0.0048994 loss) | |
| I0510 15:47:00.749979 10926 solver.cpp:245] Train net output #139: loss3/loss15 = 0.0418617 (* 0.0909091 = 0.00380561 loss) | |
| I0510 15:47:00.749994 10926 solver.cpp:245] Train net output #140: loss3/loss16 = 0.0287566 (* 0.0909091 = 0.00261424 loss) | |
| I0510 15:47:00.750008 10926 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00381576 (* 0.0909091 = 0.000346887 loss) | |
| I0510 15:47:00.750022 10926 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00244504 (* 0.0909091 = 0.000222277 loss) | |
| I0510 15:47:00.750037 10926 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00163377 (* 0.0909091 = 0.000148525 loss) | |
| I0510 15:47:00.750051 10926 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00107591 (* 0.0909091 = 9.78102e-05 loss) | |
| I0510 15:47:00.750066 10926 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00125395 (* 0.0909091 = 0.000113995 loss) | |
| I0510 15:47:00.750080 10926 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00125862 (* 0.0909091 = 0.00011442 loss) | |
| I0510 15:47:00.750093 10926 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0510 15:47:00.750104 10926 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0510 15:47:00.750116 10926 solver.cpp:245] Train net output #149: total_confidence = 1.38938e-08 | |
| I0510 15:47:00.750139 10926 solver.cpp:245] Train net output #150: total_confidence_not_rec = 3.30045e-05 | |
| I0510 15:47:00.750154 10926 sgd_solver.cpp:106] Iteration 4500, lr = 0.001 | |
| I0510 15:47:32.694175 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 30.0098 > 30) by scale factor 0.999674 | |
| I0510 15:48:34.670253 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 30.4518 > 30) by scale factor 0.985163 | |
| I0510 15:49:27.485644 10926 solver.cpp:338] Iteration 5000, Testing net (#0) | |
| I0510 15:50:10.901612 10926 solver.cpp:393] Test loss: 11.1282 | |
| I0510 15:50:10.901782 10926 solver.cpp:406] Test net output #0: loss1/accuracy = 0.0359663 | |
| I0510 15:50:10.901803 10926 solver.cpp:406] Test net output #1: loss1/accuracy01 = 0.086 | |
| I0510 15:50:10.901818 10926 solver.cpp:406] Test net output #2: loss1/accuracy02 = 0.059 | |
| I0510 15:50:10.901830 10926 solver.cpp:406] Test net output #3: loss1/accuracy03 = 0.061 | |
| I0510 15:50:10.901842 10926 solver.cpp:406] Test net output #4: loss1/accuracy04 = 0.127 | |
| I0510 15:50:10.901855 10926 solver.cpp:406] Test net output #5: loss1/accuracy05 = 0.288 | |
| I0510 15:50:10.901867 10926 solver.cpp:406] Test net output #6: loss1/accuracy06 = 0.455 | |
| I0510 15:50:10.901882 10926 solver.cpp:406] Test net output #7: loss1/accuracy07 = 0.729 | |
| I0510 15:50:10.901895 10926 solver.cpp:406] Test net output #8: loss1/accuracy08 = 0.912 | |
| I0510 15:50:10.901907 10926 solver.cpp:406] Test net output #9: loss1/accuracy09 = 0.989 | |
| I0510 15:50:10.901919 10926 solver.cpp:406] Test net output #10: loss1/accuracy10 = 0.999 | |
| I0510 15:50:10.901932 10926 solver.cpp:406] Test net output #11: loss1/accuracy11 = 1 | |
| I0510 15:50:10.901942 10926 solver.cpp:406] Test net output #12: loss1/accuracy12 = 1 | |
| I0510 15:50:10.901954 10926 solver.cpp:406] Test net output #13: loss1/accuracy13 = 1 | |
| I0510 15:50:10.901967 10926 solver.cpp:406] Test net output #14: loss1/accuracy14 = 1 | |
| I0510 15:50:10.901978 10926 solver.cpp:406] Test net output #15: loss1/accuracy15 = 1 | |
| I0510 15:50:10.901989 10926 solver.cpp:406] Test net output #16: loss1/accuracy16 = 1 | |
| I0510 15:50:10.902001 10926 solver.cpp:406] Test net output #17: loss1/accuracy17 = 1 | |
| I0510 15:50:10.902012 10926 solver.cpp:406] Test net output #18: loss1/accuracy18 = 1 | |
| I0510 15:50:10.902024 10926 solver.cpp:406] Test net output #19: loss1/accuracy19 = 1 | |
| I0510 15:50:10.902035 10926 solver.cpp:406] Test net output #20: loss1/accuracy20 = 1 | |
| I0510 15:50:10.902046 10926 solver.cpp:406] Test net output #21: loss1/accuracy21 = 1 | |
| I0510 15:50:10.902058 10926 solver.cpp:406] Test net output #22: loss1/accuracy22 = 1 | |
| I0510 15:50:10.902070 10926 solver.cpp:406] Test net output #23: loss1/accuracy_incl_empty = 0.757228 | |
| I0510 15:50:10.902088 10926 solver.cpp:406] Test net output #24: loss1/accuracy_top3 = 0.129548 | |
| I0510 15:50:10.902104 10926 solver.cpp:406] Test net output #25: loss1/cross_entropy_loss = 4.25974 (* 0.3 = 1.27792 loss) | |
| I0510 15:50:10.902119 10926 solver.cpp:406] Test net output #26: loss1/cross_entropy_loss_incl_empty = 1.11561 (* 0.3 = 0.334682 loss) | |
| I0510 15:50:10.902133 10926 solver.cpp:406] Test net output #27: loss1/loss01 = 3.65354 (* 0.0272727 = 0.099642 loss) | |
| I0510 15:50:10.902148 10926 solver.cpp:406] Test net output #28: loss1/loss02 = 3.70663 (* 0.0272727 = 0.10109 loss) | |
| I0510 15:50:10.902161 10926 solver.cpp:406] Test net output #29: loss1/loss03 = 3.74306 (* 0.0272727 = 0.102083 loss) | |
| I0510 15:50:10.902175 10926 solver.cpp:406] Test net output #30: loss1/loss04 = 3.58866 (* 0.0272727 = 0.0978724 loss) | |
| I0510 15:50:10.902189 10926 solver.cpp:406] Test net output #31: loss1/loss05 = 3.15102 (* 0.0272727 = 0.0859369 loss) | |
| I0510 15:50:10.902204 10926 solver.cpp:406] Test net output #32: loss1/loss06 = 2.63804 (* 0.0272727 = 0.0719465 loss) | |
| I0510 15:50:10.902217 10926 solver.cpp:406] Test net output #33: loss1/loss07 = 1.6046 (* 0.0272727 = 0.0437618 loss) | |
| I0510 15:50:10.902238 10926 solver.cpp:406] Test net output #34: loss1/loss08 = 0.634298 (* 0.0272727 = 0.017299 loss) | |
| I0510 15:50:10.902253 10926 solver.cpp:406] Test net output #35: loss1/loss09 = 0.146836 (* 0.0272727 = 0.00400462 loss) | |
| I0510 15:50:10.902267 10926 solver.cpp:406] Test net output #36: loss1/loss10 = 0.0794999 (* 0.0272727 = 0.00216818 loss) | |
| I0510 15:50:10.902281 10926 solver.cpp:406] Test net output #37: loss1/loss11 = 0.057249 (* 0.0272727 = 0.00156134 loss) | |
| I0510 15:50:10.902295 10926 solver.cpp:406] Test net output #38: loss1/loss12 = 0.050731 (* 0.0272727 = 0.00138357 loss) | |
| I0510 15:50:10.902317 10926 solver.cpp:406] Test net output #39: loss1/loss13 = 0.04278 (* 0.0272727 = 0.00116673 loss) | |
| I0510 15:50:10.902351 10926 solver.cpp:406] Test net output #40: loss1/loss14 = 0.0330084 (* 0.0272727 = 0.00090023 loss) | |
| I0510 15:50:10.902366 10926 solver.cpp:406] Test net output #41: loss1/loss15 = 0.0258009 (* 0.0272727 = 0.00070366 loss) | |
| I0510 15:50:10.902380 10926 solver.cpp:406] Test net output #42: loss1/loss16 = 0.0207199 (* 0.0272727 = 0.000565088 loss) | |
| I0510 15:50:10.902395 10926 solver.cpp:406] Test net output #43: loss1/loss17 = 0.0135041 (* 0.0272727 = 0.000368295 loss) | |
| I0510 15:50:10.902411 10926 solver.cpp:406] Test net output #44: loss1/loss18 = 0.0112324 (* 0.0272727 = 0.000306339 loss) | |
| I0510 15:50:10.902425 10926 solver.cpp:406] Test net output #45: loss1/loss19 = 0.00975443 (* 0.0272727 = 0.00026603 loss) | |
| I0510 15:50:10.902438 10926 solver.cpp:406] Test net output #46: loss1/loss20 = 0.00965121 (* 0.0272727 = 0.000263215 loss) | |
| I0510 15:50:10.902452 10926 solver.cpp:406] Test net output #47: loss1/loss21 = 0.0103555 (* 0.0272727 = 0.000282424 loss) | |
| I0510 15:50:10.902472 10926 solver.cpp:406] Test net output #48: loss1/loss22 = 0.00960792 (* 0.0272727 = 0.000262034 loss) | |
| I0510 15:50:10.902484 10926 solver.cpp:406] Test net output #49: loss2/accuracy = 0.0380698 | |
| I0510 15:50:10.902496 10926 solver.cpp:406] Test net output #50: loss2/accuracy01 = 0.09 | |
| I0510 15:50:10.902508 10926 solver.cpp:406] Test net output #51: loss2/accuracy02 = 0.059 | |
| I0510 15:50:10.902520 10926 solver.cpp:406] Test net output #52: loss2/accuracy03 = 0.051 | |
| I0510 15:50:10.902532 10926 solver.cpp:406] Test net output #53: loss2/accuracy04 = 0.12 | |
| I0510 15:50:10.902544 10926 solver.cpp:406] Test net output #54: loss2/accuracy05 = 0.289 | |
| I0510 15:50:10.902555 10926 solver.cpp:406] Test net output #55: loss2/accuracy06 = 0.455 | |
| I0510 15:50:10.902567 10926 solver.cpp:406] Test net output #56: loss2/accuracy07 = 0.729 | |
| I0510 15:50:10.902586 10926 solver.cpp:406] Test net output #57: loss2/accuracy08 = 0.912 | |
| I0510 15:50:10.902598 10926 solver.cpp:406] Test net output #58: loss2/accuracy09 = 0.989 | |
| I0510 15:50:10.902611 10926 solver.cpp:406] Test net output #59: loss2/accuracy10 = 0.999 | |
| I0510 15:50:10.902622 10926 solver.cpp:406] Test net output #60: loss2/accuracy11 = 1 | |
| I0510 15:50:10.902633 10926 solver.cpp:406] Test net output #61: loss2/accuracy12 = 1 | |
| I0510 15:50:10.902644 10926 solver.cpp:406] Test net output #62: loss2/accuracy13 = 1 | |
| I0510 15:50:10.902655 10926 solver.cpp:406] Test net output #63: loss2/accuracy14 = 1 | |
| I0510 15:50:10.902667 10926 solver.cpp:406] Test net output #64: loss2/accuracy15 = 1 | |
| I0510 15:50:10.902678 10926 solver.cpp:406] Test net output #65: loss2/accuracy16 = 1 | |
| I0510 15:50:10.902689 10926 solver.cpp:406] Test net output #66: loss2/accuracy17 = 1 | |
| I0510 15:50:10.902700 10926 solver.cpp:406] Test net output #67: loss2/accuracy18 = 1 | |
| I0510 15:50:10.902712 10926 solver.cpp:406] Test net output #68: loss2/accuracy19 = 1 | |
| I0510 15:50:10.902724 10926 solver.cpp:406] Test net output #69: loss2/accuracy20 = 1 | |
| I0510 15:50:10.902734 10926 solver.cpp:406] Test net output #70: loss2/accuracy21 = 1 | |
| I0510 15:50:10.902745 10926 solver.cpp:406] Test net output #71: loss2/accuracy22 = 1 | |
| I0510 15:50:10.902757 10926 solver.cpp:406] Test net output #72: loss2/accuracy_incl_empty = 0.75791 | |
| I0510 15:50:10.902770 10926 solver.cpp:406] Test net output #73: loss2/accuracy_top3 = 0.135407 | |
| I0510 15:50:10.902782 10926 solver.cpp:406] Test net output #74: loss2/cross_entropy_loss = 4.35248 (* 0.3 = 1.30574 loss) | |
| I0510 15:50:10.902801 10926 solver.cpp:406] Test net output #75: loss2/cross_entropy_loss_incl_empty = 1.14022 (* 0.3 = 0.342066 loss) | |
| I0510 15:50:10.902822 10926 solver.cpp:406] Test net output #76: loss2/loss01 = 3.64497 (* 0.0272727 = 0.0994083 loss) | |
| I0510 15:50:10.902835 10926 solver.cpp:406] Test net output #77: loss2/loss02 = 3.68193 (* 0.0272727 = 0.100416 loss) | |
| I0510 15:50:10.902848 10926 solver.cpp:406] Test net output #78: loss2/loss03 = 3.73016 (* 0.0272727 = 0.101732 loss) | |
| I0510 15:50:10.902883 10926 solver.cpp:406] Test net output #79: loss2/loss04 = 3.61512 (* 0.0272727 = 0.0985941 loss) | |
| I0510 15:50:10.902897 10926 solver.cpp:406] Test net output #80: loss2/loss05 = 3.16207 (* 0.0272727 = 0.0862384 loss) | |
| I0510 15:50:10.902911 10926 solver.cpp:406] Test net output #81: loss2/loss06 = 2.63311 (* 0.0272727 = 0.071812 loss) | |
| I0510 15:50:10.902926 10926 solver.cpp:406] Test net output #82: loss2/loss07 = 1.61077 (* 0.0272727 = 0.0439301 loss) | |
| I0510 15:50:10.902941 10926 solver.cpp:406] Test net output #83: loss2/loss08 = 0.634265 (* 0.0272727 = 0.0172981 loss) | |
| I0510 15:50:10.902956 10926 solver.cpp:406] Test net output #84: loss2/loss09 = 0.157863 (* 0.0272727 = 0.00430536 loss) | |
| I0510 15:50:10.902969 10926 solver.cpp:406] Test net output #85: loss2/loss10 = 0.0917449 (* 0.0272727 = 0.00250213 loss) | |
| I0510 15:50:10.902983 10926 solver.cpp:406] Test net output #86: loss2/loss11 = 0.0746505 (* 0.0272727 = 0.00203592 loss) | |
| I0510 15:50:10.902997 10926 solver.cpp:406] Test net output #87: loss2/loss12 = 0.0589914 (* 0.0272727 = 0.00160885 loss) | |
| I0510 15:50:10.903010 10926 solver.cpp:406] Test net output #88: loss2/loss13 = 0.0472825 (* 0.0272727 = 0.00128952 loss) | |
| I0510 15:50:10.903024 10926 solver.cpp:406] Test net output #89: loss2/loss14 = 0.0420421 (* 0.0272727 = 0.0011466 loss) | |
| I0510 15:50:10.903038 10926 solver.cpp:406] Test net output #90: loss2/loss15 = 0.0285187 (* 0.0272727 = 0.000777783 loss) | |
| I0510 15:50:10.903053 10926 solver.cpp:406] Test net output #91: loss2/loss16 = 0.0218471 (* 0.0272727 = 0.00059583 loss) | |
| I0510 15:50:10.903065 10926 solver.cpp:406] Test net output #92: loss2/loss17 = 0.0164508 (* 0.0272727 = 0.000448657 loss) | |
| I0510 15:50:10.903079 10926 solver.cpp:406] Test net output #93: loss2/loss18 = 0.0142309 (* 0.0272727 = 0.000388117 loss) | |
| I0510 15:50:10.903092 10926 solver.cpp:406] Test net output #94: loss2/loss19 = 0.0142607 (* 0.0272727 = 0.000388928 loss) | |
| I0510 15:50:10.903106 10926 solver.cpp:406] Test net output #95: loss2/loss20 = 0.0119181 (* 0.0272727 = 0.000325038 loss) | |
| I0510 15:50:10.903120 10926 solver.cpp:406] Test net output #96: loss2/loss21 = 0.0126464 (* 0.0272727 = 0.000344902 loss) | |
| I0510 15:50:10.903133 10926 solver.cpp:406] Test net output #97: loss2/loss22 = 0.0125983 (* 0.0272727 = 0.00034359 loss) | |
| I0510 15:50:10.903146 10926 solver.cpp:406] Test net output #98: loss3/accuracy = 0.0560944 | |
| I0510 15:50:10.903157 10926 solver.cpp:406] Test net output #99: loss3/accuracy01 = 0.082 | |
| I0510 15:50:10.903169 10926 solver.cpp:406] Test net output #100: loss3/accuracy02 = 0.082 | |
| I0510 15:50:10.903180 10926 solver.cpp:406] Test net output #101: loss3/accuracy03 = 0.061 | |
| I0510 15:50:10.903200 10926 solver.cpp:406] Test net output #102: loss3/accuracy04 = 0.13 | |
| I0510 15:50:10.903213 10926 solver.cpp:406] Test net output #103: loss3/accuracy05 = 0.279 | |
| I0510 15:50:10.903223 10926 solver.cpp:406] Test net output #104: loss3/accuracy06 = 0.455 | |
| I0510 15:50:10.903235 10926 solver.cpp:406] Test net output #105: loss3/accuracy07 = 0.729 | |
| I0510 15:50:10.903246 10926 solver.cpp:406] Test net output #106: loss3/accuracy08 = 0.912 | |
| I0510 15:50:10.903259 10926 solver.cpp:406] Test net output #107: loss3/accuracy09 = 0.989 | |
| I0510 15:50:10.903270 10926 solver.cpp:406] Test net output #108: loss3/accuracy10 = 0.999 | |
| I0510 15:50:10.903281 10926 solver.cpp:406] Test net output #109: loss3/accuracy11 = 1 | |
| I0510 15:50:10.903293 10926 solver.cpp:406] Test net output #110: loss3/accuracy12 = 1 | |
| I0510 15:50:10.903304 10926 solver.cpp:406] Test net output #111: loss3/accuracy13 = 1 | |
| I0510 15:50:10.903318 10926 solver.cpp:406] Test net output #112: loss3/accuracy14 = 1 | |
| I0510 15:50:10.903329 10926 solver.cpp:406] Test net output #113: loss3/accuracy15 = 1 | |
| I0510 15:50:10.903340 10926 solver.cpp:406] Test net output #114: loss3/accuracy16 = 1 | |
| I0510 15:50:10.903352 10926 solver.cpp:406] Test net output #115: loss3/accuracy17 = 1 | |
| I0510 15:50:10.903374 10926 solver.cpp:406] Test net output #116: loss3/accuracy18 = 1 | |
| I0510 15:50:10.903388 10926 solver.cpp:406] Test net output #117: loss3/accuracy19 = 1 | |
| I0510 15:50:10.903398 10926 solver.cpp:406] Test net output #118: loss3/accuracy20 = 1 | |
| I0510 15:50:10.903409 10926 solver.cpp:406] Test net output #119: loss3/accuracy21 = 1 | |
| I0510 15:50:10.903421 10926 solver.cpp:406] Test net output #120: loss3/accuracy22 = 1 | |
| I0510 15:50:10.903432 10926 solver.cpp:406] Test net output #121: loss3/accuracy_incl_empty = 0.755228 | |
| I0510 15:50:10.903445 10926 solver.cpp:406] Test net output #122: loss3/accuracy_top3 = 0.168378 | |
| I0510 15:50:10.903458 10926 solver.cpp:406] Test net output #123: loss3/cross_entropy_loss = 3.56107 (* 1 = 3.56107 loss) | |
| I0510 15:50:10.903471 10926 solver.cpp:406] Test net output #124: loss3/cross_entropy_loss_incl_empty = 1.02112 (* 1 = 1.02112 loss) | |
| I0510 15:50:10.903486 10926 solver.cpp:406] Test net output #125: loss3/loss01 = 3.43784 (* 0.0909091 = 0.312531 loss) | |
| I0510 15:50:10.903498 10926 solver.cpp:406] Test net output #126: loss3/loss02 = 3.49024 (* 0.0909091 = 0.317294 loss) | |
| I0510 15:50:10.903512 10926 solver.cpp:406] Test net output #127: loss3/loss03 = 3.55768 (* 0.0909091 = 0.323426 loss) | |
| I0510 15:50:10.903528 10926 solver.cpp:406] Test net output #128: loss3/loss04 = 3.41827 (* 0.0909091 = 0.310752 loss) | |
| I0510 15:50:10.903542 10926 solver.cpp:406] Test net output #129: loss3/loss05 = 3.00529 (* 0.0909091 = 0.273208 loss) | |
| I0510 15:50:10.903555 10926 solver.cpp:406] Test net output #130: loss3/loss06 = 2.54342 (* 0.0909091 = 0.23122 loss) | |
| I0510 15:50:10.903568 10926 solver.cpp:406] Test net output #131: loss3/loss07 = 1.5362 (* 0.0909091 = 0.139655 loss) | |
| I0510 15:50:10.903581 10926 solver.cpp:406] Test net output #132: loss3/loss08 = 0.611891 (* 0.0909091 = 0.0556264 loss) | |
| I0510 15:50:10.903600 10926 solver.cpp:406] Test net output #133: loss3/loss09 = 0.168023 (* 0.0909091 = 0.0152748 loss) | |
| I0510 15:50:10.903614 10926 solver.cpp:406] Test net output #134: loss3/loss10 = 0.105601 (* 0.0909091 = 0.00960011 loss) | |
| I0510 15:50:10.903628 10926 solver.cpp:406] Test net output #135: loss3/loss11 = 0.0714629 (* 0.0909091 = 0.00649662 loss) | |
| I0510 15:50:10.903641 10926 solver.cpp:406] Test net output #136: loss3/loss12 = 0.0635906 (* 0.0909091 = 0.00578097 loss) | |
| I0510 15:50:10.903656 10926 solver.cpp:406] Test net output #137: loss3/loss13 = 0.0507612 (* 0.0909091 = 0.00461465 loss) | |
| I0510 15:50:10.903669 10926 solver.cpp:406] Test net output #138: loss3/loss14 = 0.0392246 (* 0.0909091 = 0.00356587 loss) | |
| I0510 15:50:10.903682 10926 solver.cpp:406] Test net output #139: loss3/loss15 = 0.0283867 (* 0.0909091 = 0.00258061 loss) | |
| I0510 15:50:10.903697 10926 solver.cpp:406] Test net output #140: loss3/loss16 = 0.0185236 (* 0.0909091 = 0.00168396 loss) | |
| I0510 15:50:10.903709 10926 solver.cpp:406] Test net output #141: loss3/loss17 = 0.00822535 (* 0.0909091 = 0.000747759 loss) | |
| I0510 15:50:10.903723 10926 solver.cpp:406] Test net output #142: loss3/loss18 = 0.00528425 (* 0.0909091 = 0.000480387 loss) | |
| I0510 15:50:10.903738 10926 solver.cpp:406] Test net output #143: loss3/loss19 = 0.00507924 (* 0.0909091 = 0.000461749 loss) | |
| I0510 15:50:10.903750 10926 solver.cpp:406] Test net output #144: loss3/loss20 = 0.00391965 (* 0.0909091 = 0.000356332 loss) | |
| I0510 15:50:10.903764 10926 solver.cpp:406] Test net output #145: loss3/loss21 = 0.00283858 (* 0.0909091 = 0.000258053 loss) | |
| I0510 15:50:10.903779 10926 solver.cpp:406] Test net output #146: loss3/loss22 = 0.0021108 (* 0.0909091 = 0.000191891 loss) | |
| I0510 15:50:10.903789 10926 solver.cpp:406] Test net output #147: total_accuracy = 0 | |
| I0510 15:50:10.903801 10926 solver.cpp:406] Test net output #148: total_accuracy_not_rec = 0 | |
| I0510 15:50:10.903812 10926 solver.cpp:406] Test net output #149: total_confidence = 2.48567e-05 | |
| I0510 15:50:10.903825 10926 solver.cpp:406] Test net output #150: total_confidence_not_rec = 0.000131981 | |
| I0510 15:50:10.903851 10926 solver.cpp:338] Iteration 5000, Testing net (#1) | |
| I0510 15:50:54.547513 10926 solver.cpp:393] Test loss: 11.8769 | |
| I0510 15:50:54.547641 10926 solver.cpp:406] Test net output #0: loss1/accuracy = 0.0307114 | |
| I0510 15:50:54.547660 10926 solver.cpp:406] Test net output #1: loss1/accuracy01 = 0.097 | |
| I0510 15:50:54.547673 10926 solver.cpp:406] Test net output #2: loss1/accuracy02 = 0.072 | |
| I0510 15:50:54.547686 10926 solver.cpp:406] Test net output #3: loss1/accuracy03 = 0.055 | |
| I0510 15:50:54.547698 10926 solver.cpp:406] Test net output #4: loss1/accuracy04 = 0.107 | |
| I0510 15:50:54.547710 10926 solver.cpp:406] Test net output #5: loss1/accuracy05 = 0.247 | |
| I0510 15:50:54.547722 10926 solver.cpp:406] Test net output #6: loss1/accuracy06 = 0.401 | |
| I0510 15:50:54.547734 10926 solver.cpp:406] Test net output #7: loss1/accuracy07 = 0.66 | |
| I0510 15:50:54.547747 10926 solver.cpp:406] Test net output #8: loss1/accuracy08 = 0.814 | |
| I0510 15:50:54.547760 10926 solver.cpp:406] Test net output #9: loss1/accuracy09 = 0.897 | |
| I0510 15:50:54.547770 10926 solver.cpp:406] Test net output #10: loss1/accuracy10 = 0.92 | |
| I0510 15:50:54.547782 10926 solver.cpp:406] Test net output #11: loss1/accuracy11 = 0.937 | |
| I0510 15:50:54.547794 10926 solver.cpp:406] Test net output #12: loss1/accuracy12 = 0.945 | |
| I0510 15:50:54.547806 10926 solver.cpp:406] Test net output #13: loss1/accuracy13 = 0.959 | |
| I0510 15:50:54.547818 10926 solver.cpp:406] Test net output #14: loss1/accuracy14 = 0.97 | |
| I0510 15:50:54.547830 10926 solver.cpp:406] Test net output #15: loss1/accuracy15 = 0.977 | |
| I0510 15:50:54.547842 10926 solver.cpp:406] Test net output #16: loss1/accuracy16 = 0.984 | |
| I0510 15:50:54.547853 10926 solver.cpp:406] Test net output #17: loss1/accuracy17 = 0.994 | |
| I0510 15:50:54.547865 10926 solver.cpp:406] Test net output #18: loss1/accuracy18 = 0.996 | |
| I0510 15:50:54.547879 10926 solver.cpp:406] Test net output #19: loss1/accuracy19 = 0.997 | |
| I0510 15:50:54.547893 10926 solver.cpp:406] Test net output #20: loss1/accuracy20 = 0.997 | |
| I0510 15:50:54.547904 10926 solver.cpp:406] Test net output #21: loss1/accuracy21 = 0.998 | |
| I0510 15:50:54.547916 10926 solver.cpp:406] Test net output #22: loss1/accuracy22 = 0.999 | |
| I0510 15:50:54.547927 10926 solver.cpp:406] Test net output #23: loss1/accuracy_incl_empty = 0.725092 | |
| I0510 15:50:54.547940 10926 solver.cpp:406] Test net output #24: loss1/accuracy_top3 = 0.130296 | |
| I0510 15:50:54.547955 10926 solver.cpp:406] Test net output #25: loss1/cross_entropy_loss = 4.32543 (* 0.3 = 1.29763 loss) | |
| I0510 15:50:54.547971 10926 solver.cpp:406] Test net output #26: loss1/cross_entropy_loss_incl_empty = 1.2784 (* 0.3 = 0.38352 loss) | |
| I0510 15:50:54.547986 10926 solver.cpp:406] Test net output #27: loss1/loss01 = 3.65579 (* 0.0272727 = 0.0997035 loss) | |
| I0510 15:50:54.547999 10926 solver.cpp:406] Test net output #28: loss1/loss02 = 3.69613 (* 0.0272727 = 0.100804 loss) | |
| I0510 15:50:54.548012 10926 solver.cpp:406] Test net output #29: loss1/loss03 = 3.7991 (* 0.0272727 = 0.103612 loss) | |
| I0510 15:50:54.548027 10926 solver.cpp:406] Test net output #30: loss1/loss04 = 3.67854 (* 0.0272727 = 0.100324 loss) | |
| I0510 15:50:54.548040 10926 solver.cpp:406] Test net output #31: loss1/loss05 = 3.28719 (* 0.0272727 = 0.0896505 loss) | |
| I0510 15:50:54.548053 10926 solver.cpp:406] Test net output #32: loss1/loss06 = 2.79744 (* 0.0272727 = 0.0762939 loss) | |
| I0510 15:50:54.548068 10926 solver.cpp:406] Test net output #33: loss1/loss07 = 1.81848 (* 0.0272727 = 0.0495948 loss) | |
| I0510 15:50:54.548081 10926 solver.cpp:406] Test net output #34: loss1/loss08 = 1.0591 (* 0.0272727 = 0.0288846 loss) | |
| I0510 15:50:54.548095 10926 solver.cpp:406] Test net output #35: loss1/loss09 = 0.57466 (* 0.0272727 = 0.0156726 loss) | |
| I0510 15:50:54.548108 10926 solver.cpp:406] Test net output #36: loss1/loss10 = 0.496828 (* 0.0272727 = 0.0135499 loss) | |
| I0510 15:50:54.548122 10926 solver.cpp:406] Test net output #37: loss1/loss11 = 0.414274 (* 0.0272727 = 0.0112984 loss) | |
| I0510 15:50:54.548136 10926 solver.cpp:406] Test net output #38: loss1/loss12 = 0.372024 (* 0.0272727 = 0.0101461 loss) | |
| I0510 15:50:54.548151 10926 solver.cpp:406] Test net output #39: loss1/loss13 = 0.286704 (* 0.0272727 = 0.00781921 loss) | |
| I0510 15:50:54.548184 10926 solver.cpp:406] Test net output #40: loss1/loss14 = 0.227522 (* 0.0272727 = 0.00620514 loss) | |
| I0510 15:50:54.548200 10926 solver.cpp:406] Test net output #41: loss1/loss15 = 0.179006 (* 0.0272727 = 0.00488197 loss) | |
| I0510 15:50:54.548214 10926 solver.cpp:406] Test net output #42: loss1/loss16 = 0.135973 (* 0.0272727 = 0.00370836 loss) | |
| I0510 15:50:54.548228 10926 solver.cpp:406] Test net output #43: loss1/loss17 = 0.0600078 (* 0.0272727 = 0.00163658 loss) | |
| I0510 15:50:54.548243 10926 solver.cpp:406] Test net output #44: loss1/loss18 = 0.0454324 (* 0.0272727 = 0.00123906 loss) | |
| I0510 15:50:54.548256 10926 solver.cpp:406] Test net output #45: loss1/loss19 = 0.0348575 (* 0.0272727 = 0.00095066 loss) | |
| I0510 15:50:54.548270 10926 solver.cpp:406] Test net output #46: loss1/loss20 = 0.0351693 (* 0.0272727 = 0.000959163 loss) | |
| I0510 15:50:54.548285 10926 solver.cpp:406] Test net output #47: loss1/loss21 = 0.0278049 (* 0.0272727 = 0.000758315 loss) | |
| I0510 15:50:54.548298 10926 solver.cpp:406] Test net output #48: loss1/loss22 = 0.0197249 (* 0.0272727 = 0.000537951 loss) | |
| I0510 15:50:54.548310 10926 solver.cpp:406] Test net output #49: loss2/accuracy = 0.0312412 | |
| I0510 15:50:54.548322 10926 solver.cpp:406] Test net output #50: loss2/accuracy01 = 0.099 | |
| I0510 15:50:54.548334 10926 solver.cpp:406] Test net output #51: loss2/accuracy02 = 0.064 | |
| I0510 15:50:54.548346 10926 solver.cpp:406] Test net output #52: loss2/accuracy03 = 0.046 | |
| I0510 15:50:54.548357 10926 solver.cpp:406] Test net output #53: loss2/accuracy04 = 0.098 | |
| I0510 15:50:54.548369 10926 solver.cpp:406] Test net output #54: loss2/accuracy05 = 0.247 | |
| I0510 15:50:54.548382 10926 solver.cpp:406] Test net output #55: loss2/accuracy06 = 0.401 | |
| I0510 15:50:54.548393 10926 solver.cpp:406] Test net output #56: loss2/accuracy07 = 0.66 | |
| I0510 15:50:54.548404 10926 solver.cpp:406] Test net output #57: loss2/accuracy08 = 0.814 | |
| I0510 15:50:54.548415 10926 solver.cpp:406] Test net output #58: loss2/accuracy09 = 0.897 | |
| I0510 15:50:54.548427 10926 solver.cpp:406] Test net output #59: loss2/accuracy10 = 0.92 | |
| I0510 15:50:54.548439 10926 solver.cpp:406] Test net output #60: loss2/accuracy11 = 0.937 | |
| I0510 15:50:54.548450 10926 solver.cpp:406] Test net output #61: loss2/accuracy12 = 0.945 | |
| I0510 15:50:54.548461 10926 solver.cpp:406] Test net output #62: loss2/accuracy13 = 0.959 | |
| I0510 15:50:54.548473 10926 solver.cpp:406] Test net output #63: loss2/accuracy14 = 0.97 | |
| I0510 15:50:54.548485 10926 solver.cpp:406] Test net output #64: loss2/accuracy15 = 0.977 | |
| I0510 15:50:54.548496 10926 solver.cpp:406] Test net output #65: loss2/accuracy16 = 0.984 | |
| I0510 15:50:54.548507 10926 solver.cpp:406] Test net output #66: loss2/accuracy17 = 0.994 | |
| I0510 15:50:54.548518 10926 solver.cpp:406] Test net output #67: loss2/accuracy18 = 0.996 | |
| I0510 15:50:54.548530 10926 solver.cpp:406] Test net output #68: loss2/accuracy19 = 0.997 | |
| I0510 15:50:54.548542 10926 solver.cpp:406] Test net output #69: loss2/accuracy20 = 0.997 | |
| I0510 15:50:54.548553 10926 solver.cpp:406] Test net output #70: loss2/accuracy21 = 0.998 | |
| I0510 15:50:54.548564 10926 solver.cpp:406] Test net output #71: loss2/accuracy22 = 0.999 | |
| I0510 15:50:54.548576 10926 solver.cpp:406] Test net output #72: loss2/accuracy_incl_empty = 0.725319 | |
| I0510 15:50:54.548588 10926 solver.cpp:406] Test net output #73: loss2/accuracy_top3 = 0.129488 | |
| I0510 15:50:54.548601 10926 solver.cpp:406] Test net output #74: loss2/cross_entropy_loss = 4.43663 (* 0.3 = 1.33099 loss) | |
| I0510 15:50:54.548615 10926 solver.cpp:406] Test net output #75: loss2/cross_entropy_loss_incl_empty = 1.31661 (* 0.3 = 0.394983 loss) | |
| I0510 15:50:54.548629 10926 solver.cpp:406] Test net output #76: loss2/loss01 = 3.65947 (* 0.0272727 = 0.0998037 loss) | |
| I0510 15:50:54.548642 10926 solver.cpp:406] Test net output #77: loss2/loss02 = 3.68544 (* 0.0272727 = 0.100512 loss) | |
| I0510 15:50:54.548671 10926 solver.cpp:406] Test net output #78: loss2/loss03 = 3.7855 (* 0.0272727 = 0.103241 loss) | |
| I0510 15:50:54.548686 10926 solver.cpp:406] Test net output #79: loss2/loss04 = 3.68562 (* 0.0272727 = 0.100517 loss) | |
| I0510 15:50:54.548701 10926 solver.cpp:406] Test net output #80: loss2/loss05 = 3.29085 (* 0.0272727 = 0.0897505 loss) | |
| I0510 15:50:54.548714 10926 solver.cpp:406] Test net output #81: loss2/loss06 = 2.80489 (* 0.0272727 = 0.0764971 loss) | |
| I0510 15:50:54.548727 10926 solver.cpp:406] Test net output #82: loss2/loss07 = 1.83156 (* 0.0272727 = 0.0499515 loss) | |
| I0510 15:50:54.548740 10926 solver.cpp:406] Test net output #83: loss2/loss08 = 1.05517 (* 0.0272727 = 0.0287774 loss) | |
| I0510 15:50:54.548753 10926 solver.cpp:406] Test net output #84: loss2/loss09 = 0.600013 (* 0.0272727 = 0.016364 loss) | |
| I0510 15:50:54.548768 10926 solver.cpp:406] Test net output #85: loss2/loss10 = 0.497504 (* 0.0272727 = 0.0135683 loss) | |
| I0510 15:50:54.548780 10926 solver.cpp:406] Test net output #86: loss2/loss11 = 0.425039 (* 0.0272727 = 0.011592 loss) | |
| I0510 15:50:54.548794 10926 solver.cpp:406] Test net output #87: loss2/loss12 = 0.385258 (* 0.0272727 = 0.010507 loss) | |
| I0510 15:50:54.548809 10926 solver.cpp:406] Test net output #88: loss2/loss13 = 0.294978 (* 0.0272727 = 0.00804485 loss) | |
| I0510 15:50:54.548821 10926 solver.cpp:406] Test net output #89: loss2/loss14 = 0.232285 (* 0.0272727 = 0.00633505 loss) | |
| I0510 15:50:54.548835 10926 solver.cpp:406] Test net output #90: loss2/loss15 = 0.18327 (* 0.0272727 = 0.00499828 loss) | |
| I0510 15:50:54.548848 10926 solver.cpp:406] Test net output #91: loss2/loss16 = 0.139492 (* 0.0272727 = 0.00380432 loss) | |
| I0510 15:50:54.548862 10926 solver.cpp:406] Test net output #92: loss2/loss17 = 0.0632308 (* 0.0272727 = 0.00172448 loss) | |
| I0510 15:50:54.548877 10926 solver.cpp:406] Test net output #93: loss2/loss18 = 0.0474037 (* 0.0272727 = 0.00129283 loss) | |
| I0510 15:50:54.548889 10926 solver.cpp:406] Test net output #94: loss2/loss19 = 0.0397462 (* 0.0272727 = 0.00108399 loss) | |
| I0510 15:50:54.548902 10926 solver.cpp:406] Test net output #95: loss2/loss20 = 0.0377938 (* 0.0272727 = 0.00103074 loss) | |
| I0510 15:50:54.548916 10926 solver.cpp:406] Test net output #96: loss2/loss21 = 0.0310736 (* 0.0272727 = 0.000847461 loss) | |
| I0510 15:50:54.548931 10926 solver.cpp:406] Test net output #97: loss2/loss22 = 0.0229608 (* 0.0272727 = 0.000626204 loss) | |
| I0510 15:50:54.548944 10926 solver.cpp:406] Test net output #98: loss3/accuracy = 0.0619427 | |
| I0510 15:50:54.548955 10926 solver.cpp:406] Test net output #99: loss3/accuracy01 = 0.1 | |
| I0510 15:50:54.548967 10926 solver.cpp:406] Test net output #100: loss3/accuracy02 = 0.081 | |
| I0510 15:50:54.548979 10926 solver.cpp:406] Test net output #101: loss3/accuracy03 = 0.056 | |
| I0510 15:50:54.548990 10926 solver.cpp:406] Test net output #102: loss3/accuracy04 = 0.113 | |
| I0510 15:50:54.549002 10926 solver.cpp:406] Test net output #103: loss3/accuracy05 = 0.25 | |
| I0510 15:50:54.549013 10926 solver.cpp:406] Test net output #104: loss3/accuracy06 = 0.401 | |
| I0510 15:50:54.549026 10926 solver.cpp:406] Test net output #105: loss3/accuracy07 = 0.66 | |
| I0510 15:50:54.549036 10926 solver.cpp:406] Test net output #106: loss3/accuracy08 = 0.814 | |
| I0510 15:50:54.549047 10926 solver.cpp:406] Test net output #107: loss3/accuracy09 = 0.897 | |
| I0510 15:50:54.549059 10926 solver.cpp:406] Test net output #108: loss3/accuracy10 = 0.92 | |
| I0510 15:50:54.549070 10926 solver.cpp:406] Test net output #109: loss3/accuracy11 = 0.937 | |
| I0510 15:50:54.549082 10926 solver.cpp:406] Test net output #110: loss3/accuracy12 = 0.945 | |
| I0510 15:50:54.549093 10926 solver.cpp:406] Test net output #111: loss3/accuracy13 = 0.959 | |
| I0510 15:50:54.549101 10926 solver.cpp:406] Test net output #112: loss3/accuracy14 = 0.97 | |
| I0510 15:50:54.549109 10926 solver.cpp:406] Test net output #113: loss3/accuracy15 = 0.977 | |
| I0510 15:50:54.549129 10926 solver.cpp:406] Test net output #114: loss3/accuracy16 = 0.984 | |
| I0510 15:50:54.549154 10926 solver.cpp:406] Test net output #115: loss3/accuracy17 = 0.994 | |
| I0510 15:50:54.549167 10926 solver.cpp:406] Test net output #116: loss3/accuracy18 = 0.996 | |
| I0510 15:50:54.549180 10926 solver.cpp:406] Test net output #117: loss3/accuracy19 = 0.997 | |
| I0510 15:50:54.549190 10926 solver.cpp:406] Test net output #118: loss3/accuracy20 = 0.997 | |
| I0510 15:50:54.549202 10926 solver.cpp:406] Test net output #119: loss3/accuracy21 = 0.998 | |
| I0510 15:50:54.549213 10926 solver.cpp:406] Test net output #120: loss3/accuracy22 = 0.999 | |
| I0510 15:50:54.549224 10926 solver.cpp:406] Test net output #121: loss3/accuracy_incl_empty = 0.727546 | |
| I0510 15:50:54.549237 10926 solver.cpp:406] Test net output #122: loss3/accuracy_top3 = 0.179784 | |
| I0510 15:50:54.549249 10926 solver.cpp:406] Test net output #123: loss3/cross_entropy_loss = 3.57406 (* 1 = 3.57406 loss) | |
| I0510 15:50:54.549263 10926 solver.cpp:406] Test net output #124: loss3/cross_entropy_loss_incl_empty = 1.14621 (* 1 = 1.14621 loss) | |
| I0510 15:50:54.549276 10926 solver.cpp:406] Test net output #125: loss3/loss01 = 3.40776 (* 0.0909091 = 0.309797 loss) | |
| I0510 15:50:54.549289 10926 solver.cpp:406] Test net output #126: loss3/loss02 = 3.47026 (* 0.0909091 = 0.315478 loss) | |
| I0510 15:50:54.549302 10926 solver.cpp:406] Test net output #127: loss3/loss03 = 3.56049 (* 0.0909091 = 0.323681 loss) | |
| I0510 15:50:54.549315 10926 solver.cpp:406] Test net output #128: loss3/loss04 = 3.5185 (* 0.0909091 = 0.319864 loss) | |
| I0510 15:50:54.549329 10926 solver.cpp:406] Test net output #129: loss3/loss05 = 3.16383 (* 0.0909091 = 0.287621 loss) | |
| I0510 15:50:54.549341 10926 solver.cpp:406] Test net output #130: loss3/loss06 = 2.69373 (* 0.0909091 = 0.244884 loss) | |
| I0510 15:50:54.549355 10926 solver.cpp:406] Test net output #131: loss3/loss07 = 1.71829 (* 0.0909091 = 0.156209 loss) | |
| I0510 15:50:54.549368 10926 solver.cpp:406] Test net output #132: loss3/loss08 = 0.979792 (* 0.0909091 = 0.089072 loss) | |
| I0510 15:50:54.549381 10926 solver.cpp:406] Test net output #133: loss3/loss09 = 0.540733 (* 0.0909091 = 0.0491575 loss) | |
| I0510 15:50:54.549394 10926 solver.cpp:406] Test net output #134: loss3/loss10 = 0.461482 (* 0.0909091 = 0.0419529 loss) | |
| I0510 15:50:54.549408 10926 solver.cpp:406] Test net output #135: loss3/loss11 = 0.378684 (* 0.0909091 = 0.0344258 loss) | |
| I0510 15:50:54.549422 10926 solver.cpp:406] Test net output #136: loss3/loss12 = 0.35143 (* 0.0909091 = 0.0319482 loss) | |
| I0510 15:50:54.549434 10926 solver.cpp:406] Test net output #137: loss3/loss13 = 0.258869 (* 0.0909091 = 0.0235336 loss) | |
| I0510 15:50:54.549448 10926 solver.cpp:406] Test net output #138: loss3/loss14 = 0.20106 (* 0.0909091 = 0.0182782 loss) | |
| I0510 15:50:54.549461 10926 solver.cpp:406] Test net output #139: loss3/loss15 = 0.160705 (* 0.0909091 = 0.0146096 loss) | |
| I0510 15:50:54.549474 10926 solver.cpp:406] Test net output #140: loss3/loss16 = 0.123306 (* 0.0909091 = 0.0112097 loss) | |
| I0510 15:50:54.549487 10926 solver.cpp:406] Test net output #141: loss3/loss17 = 0.0567447 (* 0.0909091 = 0.00515861 loss) | |
| I0510 15:50:54.549501 10926 solver.cpp:406] Test net output #142: loss3/loss18 = 0.0438257 (* 0.0909091 = 0.00398416 loss) | |
| I0510 15:50:54.549515 10926 solver.cpp:406] Test net output #143: loss3/loss19 = 0.0337243 (* 0.0909091 = 0.00306585 loss) | |
| I0510 15:50:54.549528 10926 solver.cpp:406] Test net output #144: loss3/loss20 = 0.034632 (* 0.0909091 = 0.00314836 loss) | |
| I0510 15:50:54.549541 10926 solver.cpp:406] Test net output #145: loss3/loss21 = 0.0223054 (* 0.0909091 = 0.00202777 loss) | |
| I0510 15:50:54.549556 10926 solver.cpp:406] Test net output #146: loss3/loss22 = 0.01385 (* 0.0909091 = 0.00125909 loss) | |
| I0510 15:50:54.549567 10926 solver.cpp:406] Test net output #147: total_accuracy = 0 | |
| I0510 15:50:54.549578 10926 solver.cpp:406] Test net output #148: total_accuracy_not_rec = 0 | |
| I0510 15:50:54.549588 10926 solver.cpp:406] Test net output #149: total_confidence = 1.88937e-05 | |
| I0510 15:50:54.549610 10926 solver.cpp:406] Test net output #150: total_confidence_not_rec = 9.85406e-05 | |
| I0510 15:50:54.694993 10926 solver.cpp:229] Iteration 5000, loss = 11.9024 | |
| I0510 15:50:54.695055 10926 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0243902 | |
| I0510 15:50:54.695072 10926 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0510 15:50:54.695086 10926 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0510 15:50:54.695098 10926 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.125 | |
| I0510 15:50:54.695111 10926 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.125 | |
| I0510 15:50:54.695123 10926 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.375 | |
| I0510 15:50:54.695137 10926 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.625 | |
| I0510 15:50:54.695148 10926 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.875 | |
| I0510 15:50:54.695161 10926 solver.cpp:245] Train net output #8: loss1/accuracy08 = 1 | |
| I0510 15:50:54.695176 10926 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0510 15:50:54.695189 10926 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0510 15:50:54.695202 10926 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0510 15:50:54.695214 10926 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0510 15:50:54.695226 10926 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0510 15:50:54.695238 10926 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0510 15:50:54.695250 10926 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0510 15:50:54.695262 10926 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0510 15:50:54.695274 10926 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0510 15:50:54.695286 10926 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0510 15:50:54.695298 10926 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0510 15:50:54.695309 10926 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0510 15:50:54.695322 10926 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0510 15:50:54.695333 10926 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0510 15:50:54.695344 10926 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.744318 | |
| I0510 15:50:54.695356 10926 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.097561 | |
| I0510 15:50:54.695374 10926 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.68148 (* 0.3 = 1.10444 loss) | |
| I0510 15:50:54.695389 10926 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.28626 (* 0.3 = 0.385879 loss) | |
| I0510 15:50:54.695402 10926 solver.cpp:245] Train net output #27: loss1/loss01 = 3.75503 (* 0.0272727 = 0.10241 loss) | |
| I0510 15:50:54.695416 10926 solver.cpp:245] Train net output #28: loss1/loss02 = 3.73541 (* 0.0272727 = 0.101875 loss) | |
| I0510 15:50:54.695430 10926 solver.cpp:245] Train net output #29: loss1/loss03 = 3.37046 (* 0.0272727 = 0.0919215 loss) | |
| I0510 15:50:54.695446 10926 solver.cpp:245] Train net output #30: loss1/loss04 = 3.71733 (* 0.0272727 = 0.101382 loss) | |
| I0510 15:50:54.695459 10926 solver.cpp:245] Train net output #31: loss1/loss05 = 3.08925 (* 0.0272727 = 0.0842523 loss) | |
| I0510 15:50:54.695473 10926 solver.cpp:245] Train net output #32: loss1/loss06 = 2.30991 (* 0.0272727 = 0.0629976 loss) | |
| I0510 15:50:54.695487 10926 solver.cpp:245] Train net output #33: loss1/loss07 = 1.0563 (* 0.0272727 = 0.0288082 loss) | |
| I0510 15:50:54.695502 10926 solver.cpp:245] Train net output #34: loss1/loss08 = 0.381119 (* 0.0272727 = 0.0103942 loss) | |
| I0510 15:50:54.695516 10926 solver.cpp:245] Train net output #35: loss1/loss09 = 0.191784 (* 0.0272727 = 0.00523047 loss) | |
| I0510 15:50:54.695538 10926 solver.cpp:245] Train net output #36: loss1/loss10 = 0.128585 (* 0.0272727 = 0.00350686 loss) | |
| I0510 15:50:54.695566 10926 solver.cpp:245] Train net output #37: loss1/loss11 = 0.11433 (* 0.0272727 = 0.00311808 loss) | |
| I0510 15:50:54.695623 10926 solver.cpp:245] Train net output #38: loss1/loss12 = 0.132261 (* 0.0272727 = 0.00360712 loss) | |
| I0510 15:50:54.695641 10926 solver.cpp:245] Train net output #39: loss1/loss13 = 0.0674421 (* 0.0272727 = 0.00183933 loss) | |
| I0510 15:50:54.695655 10926 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0679602 (* 0.0272727 = 0.00185346 loss) | |
| I0510 15:50:54.695670 10926 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0485562 (* 0.0272727 = 0.00132426 loss) | |
| I0510 15:50:54.695688 10926 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0434146 (* 0.0272727 = 0.00118404 loss) | |
| I0510 15:50:54.695703 10926 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0227302 (* 0.0272727 = 0.000619914 loss) | |
| I0510 15:50:54.695719 10926 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0356345 (* 0.0272727 = 0.000971851 loss) | |
| I0510 15:50:54.695734 10926 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0241349 (* 0.0272727 = 0.000658225 loss) | |
| I0510 15:50:54.695747 10926 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0241058 (* 0.0272727 = 0.000657431 loss) | |
| I0510 15:50:54.695762 10926 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0246037 (* 0.0272727 = 0.00067101 loss) | |
| I0510 15:50:54.695776 10926 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0190219 (* 0.0272727 = 0.000518779 loss) | |
| I0510 15:50:54.695790 10926 solver.cpp:245] Train net output #49: loss2/accuracy = 0 | |
| I0510 15:50:54.695801 10926 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.125 | |
| I0510 15:50:54.695814 10926 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0.125 | |
| I0510 15:50:54.695827 10926 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0.25 | |
| I0510 15:50:54.695838 10926 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0 | |
| I0510 15:50:54.695850 10926 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.375 | |
| I0510 15:50:54.695863 10926 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.625 | |
| I0510 15:50:54.695874 10926 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.875 | |
| I0510 15:50:54.695886 10926 solver.cpp:245] Train net output #57: loss2/accuracy08 = 1 | |
| I0510 15:50:54.695899 10926 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0510 15:50:54.695910 10926 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0510 15:50:54.695921 10926 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0510 15:50:54.695933 10926 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0510 15:50:54.695945 10926 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0510 15:50:54.695957 10926 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0510 15:50:54.695968 10926 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0510 15:50:54.695981 10926 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0510 15:50:54.695992 10926 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0510 15:50:54.696003 10926 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0510 15:50:54.696015 10926 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0510 15:50:54.696027 10926 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0510 15:50:54.696038 10926 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0510 15:50:54.696049 10926 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0510 15:50:54.696061 10926 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.761364 | |
| I0510 15:50:54.696074 10926 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.0731707 | |
| I0510 15:50:54.696089 10926 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.66806 (* 0.3 = 1.10042 loss) | |
| I0510 15:50:54.696102 10926 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.09451 (* 0.3 = 0.328354 loss) | |
| I0510 15:50:54.696117 10926 solver.cpp:245] Train net output #76: loss2/loss01 = 3.84036 (* 0.0272727 = 0.104737 loss) | |
| I0510 15:50:54.696144 10926 solver.cpp:245] Train net output #77: loss2/loss02 = 3.79634 (* 0.0272727 = 0.103537 loss) | |
| I0510 15:50:54.696159 10926 solver.cpp:245] Train net output #78: loss2/loss03 = 3.15302 (* 0.0272727 = 0.0859914 loss) | |
| I0510 15:50:54.696173 10926 solver.cpp:245] Train net output #79: loss2/loss04 = 3.78145 (* 0.0272727 = 0.10313 loss) | |
| I0510 15:50:54.696188 10926 solver.cpp:245] Train net output #80: loss2/loss05 = 3.19878 (* 0.0272727 = 0.0872394 loss) | |
| I0510 15:50:54.696202 10926 solver.cpp:245] Train net output #81: loss2/loss06 = 2.0117 (* 0.0272727 = 0.0548646 loss) | |
| I0510 15:50:54.696219 10926 solver.cpp:245] Train net output #82: loss2/loss07 = 0.834964 (* 0.0272727 = 0.0227718 loss) | |
| I0510 15:50:54.696235 10926 solver.cpp:245] Train net output #83: loss2/loss08 = 0.263192 (* 0.0272727 = 0.00717796 loss) | |
| I0510 15:50:54.696250 10926 solver.cpp:245] Train net output #84: loss2/loss09 = 0.159556 (* 0.0272727 = 0.00435154 loss) | |
| I0510 15:50:54.696264 10926 solver.cpp:245] Train net output #85: loss2/loss10 = 0.170671 (* 0.0272727 = 0.00465466 loss) | |
| I0510 15:50:54.696279 10926 solver.cpp:245] Train net output #86: loss2/loss11 = 0.0745095 (* 0.0272727 = 0.00203208 loss) | |
| I0510 15:50:54.696293 10926 solver.cpp:245] Train net output #87: loss2/loss12 = 0.0735788 (* 0.0272727 = 0.00200669 loss) | |
| I0510 15:50:54.696307 10926 solver.cpp:245] Train net output #88: loss2/loss13 = 0.0560863 (* 0.0272727 = 0.00152963 loss) | |
| I0510 15:50:54.696321 10926 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0513314 (* 0.0272727 = 0.00139995 loss) | |
| I0510 15:50:54.696334 10926 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0522384 (* 0.0272727 = 0.00142468 loss) | |
| I0510 15:50:54.696349 10926 solver.cpp:245] Train net output #91: loss2/loss16 = 0.023128 (* 0.0272727 = 0.000630762 loss) | |
| I0510 15:50:54.696363 10926 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0246731 (* 0.0272727 = 0.000672902 loss) | |
| I0510 15:50:54.696377 10926 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0258979 (* 0.0272727 = 0.000706306 loss) | |
| I0510 15:50:54.696391 10926 solver.cpp:245] Train net output #94: loss2/loss19 = 0.0193419 (* 0.0272727 = 0.000527508 loss) | |
| I0510 15:50:54.696404 10926 solver.cpp:245] Train net output #95: loss2/loss20 = 0.0192978 (* 0.0272727 = 0.000526303 loss) | |
| I0510 15:50:54.696419 10926 solver.cpp:245] Train net output #96: loss2/loss21 = 0.0113269 (* 0.0272727 = 0.000308915 loss) | |
| I0510 15:50:54.696432 10926 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0141585 (* 0.0272727 = 0.000386142 loss) | |
| I0510 15:50:54.696444 10926 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0731707 | |
| I0510 15:50:54.696457 10926 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0510 15:50:54.696470 10926 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0510 15:50:54.696480 10926 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0510 15:50:54.696492 10926 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0510 15:50:54.696504 10926 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.375 | |
| I0510 15:50:54.696516 10926 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.625 | |
| I0510 15:50:54.696527 10926 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.875 | |
| I0510 15:50:54.696539 10926 solver.cpp:245] Train net output #106: loss3/accuracy08 = 1 | |
| I0510 15:50:54.696550 10926 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0510 15:50:54.696563 10926 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0510 15:50:54.696574 10926 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0510 15:50:54.696585 10926 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0510 15:50:54.696598 10926 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0510 15:50:54.696609 10926 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0510 15:50:54.696631 10926 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0510 15:50:54.696645 10926 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0510 15:50:54.696656 10926 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0510 15:50:54.696667 10926 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0510 15:50:54.696679 10926 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0510 15:50:54.696691 10926 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0510 15:50:54.696703 10926 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0510 15:50:54.696715 10926 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0510 15:50:54.696727 10926 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.732955 | |
| I0510 15:50:54.696741 10926 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.243902 | |
| I0510 15:50:54.696756 10926 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.16367 (* 1 = 3.16367 loss) | |
| I0510 15:50:54.696770 10926 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.18118 (* 1 = 1.18118 loss) | |
| I0510 15:50:54.696784 10926 solver.cpp:245] Train net output #125: loss3/loss01 = 3.25279 (* 0.0909091 = 0.295708 loss) | |
| I0510 15:50:54.696799 10926 solver.cpp:245] Train net output #126: loss3/loss02 = 3.29622 (* 0.0909091 = 0.299656 loss) | |
| I0510 15:50:54.696812 10926 solver.cpp:245] Train net output #127: loss3/loss03 = 3.35421 (* 0.0909091 = 0.304928 loss) | |
| I0510 15:50:54.696826 10926 solver.cpp:245] Train net output #128: loss3/loss04 = 3.71438 (* 0.0909091 = 0.33767 loss) | |
| I0510 15:50:54.696841 10926 solver.cpp:245] Train net output #129: loss3/loss05 = 3.06186 (* 0.0909091 = 0.278351 loss) | |
| I0510 15:50:54.696851 10926 solver.cpp:245] Train net output #130: loss3/loss06 = 1.84348 (* 0.0909091 = 0.167589 loss) | |
| I0510 15:50:54.696866 10926 solver.cpp:245] Train net output #131: loss3/loss07 = 0.881619 (* 0.0909091 = 0.0801472 loss) | |
| I0510 15:50:54.696879 10926 solver.cpp:245] Train net output #132: loss3/loss08 = 0.430755 (* 0.0909091 = 0.0391596 loss) | |
| I0510 15:50:54.696893 10926 solver.cpp:245] Train net output #133: loss3/loss09 = 0.297638 (* 0.0909091 = 0.027058 loss) | |
| I0510 15:50:54.696907 10926 solver.cpp:245] Train net output #134: loss3/loss10 = 0.197644 (* 0.0909091 = 0.0179676 loss) | |
| I0510 15:50:54.696921 10926 solver.cpp:245] Train net output #135: loss3/loss11 = 0.142417 (* 0.0909091 = 0.012947 loss) | |
| I0510 15:50:54.696935 10926 solver.cpp:245] Train net output #136: loss3/loss12 = 0.165868 (* 0.0909091 = 0.0150789 loss) | |
| I0510 15:50:54.696949 10926 solver.cpp:245] Train net output #137: loss3/loss13 = 0.0808731 (* 0.0909091 = 0.0073521 loss) | |
| I0510 15:50:54.696964 10926 solver.cpp:245] Train net output #138: loss3/loss14 = 0.0876784 (* 0.0909091 = 0.00797076 loss) | |
| I0510 15:50:54.696979 10926 solver.cpp:245] Train net output #139: loss3/loss15 = 0.0322726 (* 0.0909091 = 0.00293387 loss) | |
| I0510 15:50:54.696991 10926 solver.cpp:245] Train net output #140: loss3/loss16 = 0.023687 (* 0.0909091 = 0.00215337 loss) | |
| I0510 15:50:54.697005 10926 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00482418 (* 0.0909091 = 0.000438562 loss) | |
| I0510 15:50:54.697019 10926 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00245039 (* 0.0909091 = 0.000222763 loss) | |
| I0510 15:50:54.697033 10926 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00126022 (* 0.0909091 = 0.000114566 loss) | |
| I0510 15:50:54.697047 10926 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00120991 (* 0.0909091 = 0.000109992 loss) | |
| I0510 15:50:54.697062 10926 solver.cpp:245] Train net output #145: loss3/loss21 = 0.000965282 (* 0.0909091 = 8.77529e-05 loss) | |
| I0510 15:50:54.697074 10926 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00093468 (* 0.0909091 = 8.49709e-05 loss) | |
| I0510 15:50:54.697088 10926 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0510 15:50:54.697108 10926 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0510 15:50:54.697136 10926 solver.cpp:245] Train net output #149: total_confidence = 7.04843e-06 | |
| I0510 15:50:54.697150 10926 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.000811952 | |
| I0510 15:50:54.697165 10926 sgd_solver.cpp:106] Iteration 5000, lr = 0.001 | |
| I0510 15:50:59.909981 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 31.4155 > 30) by scale factor 0.954943 | |
| I0510 15:51:25.928398 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 32.2427 > 30) by scale factor 0.930442 | |
| I0510 15:51:35.680492 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 36.9913 > 30) by scale factor 0.811001 | |
| I0510 15:52:19.668812 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 32.342 > 30) by scale factor 0.927587 | |
| I0510 15:52:22.611356 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 36.2797 > 30) by scale factor 0.826909 | |
| I0510 15:53:22.444916 10926 solver.cpp:229] Iteration 5500, loss = 11.677 | |
| I0510 15:53:22.445044 10926 solver.cpp:245] Train net output #0: loss1/accuracy = 0 | |
| I0510 15:53:22.445065 10926 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0510 15:53:22.445077 10926 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.25 | |
| I0510 15:53:22.445091 10926 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.125 | |
| I0510 15:53:22.445103 10926 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0 | |
| I0510 15:53:22.445116 10926 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.25 | |
| I0510 15:53:22.445128 10926 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.375 | |
| I0510 15:53:22.445140 10926 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.625 | |
| I0510 15:53:22.445153 10926 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.75 | |
| I0510 15:53:22.445180 10926 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.875 | |
| I0510 15:53:22.445195 10926 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.875 | |
| I0510 15:53:22.445209 10926 solver.cpp:245] Train net output #11: loss1/accuracy11 = 0.875 | |
| I0510 15:53:22.445221 10926 solver.cpp:245] Train net output #12: loss1/accuracy12 = 0.875 | |
| I0510 15:53:22.445233 10926 solver.cpp:245] Train net output #13: loss1/accuracy13 = 0.875 | |
| I0510 15:53:22.445245 10926 solver.cpp:245] Train net output #14: loss1/accuracy14 = 0.875 | |
| I0510 15:53:22.445258 10926 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0510 15:53:22.445271 10926 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0510 15:53:22.445282 10926 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0510 15:53:22.445293 10926 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0510 15:53:22.445305 10926 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0510 15:53:22.445317 10926 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0510 15:53:22.445329 10926 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0510 15:53:22.445340 10926 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0510 15:53:22.445353 10926 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.681818 | |
| I0510 15:53:22.445364 10926 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.0943396 | |
| I0510 15:53:22.445382 10926 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 4.0976 (* 0.3 = 1.22928 loss) | |
| I0510 15:53:22.445397 10926 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.53968 (* 0.3 = 0.461904 loss) | |
| I0510 15:53:22.445412 10926 solver.cpp:245] Train net output #27: loss1/loss01 = 4.03667 (* 0.0272727 = 0.110091 loss) | |
| I0510 15:53:22.445426 10926 solver.cpp:245] Train net output #28: loss1/loss02 = 3.68561 (* 0.0272727 = 0.100517 loss) | |
| I0510 15:53:22.445441 10926 solver.cpp:245] Train net output #29: loss1/loss03 = 3.45118 (* 0.0272727 = 0.0941232 loss) | |
| I0510 15:53:22.445456 10926 solver.cpp:245] Train net output #30: loss1/loss04 = 4.44707 (* 0.0272727 = 0.121284 loss) | |
| I0510 15:53:22.445469 10926 solver.cpp:245] Train net output #31: loss1/loss05 = 3.3116 (* 0.0272727 = 0.0903163 loss) | |
| I0510 15:53:22.445483 10926 solver.cpp:245] Train net output #32: loss1/loss06 = 3.04595 (* 0.0272727 = 0.0830715 loss) | |
| I0510 15:53:22.445499 10926 solver.cpp:245] Train net output #33: loss1/loss07 = 2.24527 (* 0.0272727 = 0.0612346 loss) | |
| I0510 15:53:22.445513 10926 solver.cpp:245] Train net output #34: loss1/loss08 = 1.7152 (* 0.0272727 = 0.0467781 loss) | |
| I0510 15:53:22.445528 10926 solver.cpp:245] Train net output #35: loss1/loss09 = 0.540517 (* 0.0272727 = 0.0147414 loss) | |
| I0510 15:53:22.445541 10926 solver.cpp:245] Train net output #36: loss1/loss10 = 1.01711 (* 0.0272727 = 0.0277392 loss) | |
| I0510 15:53:22.445555 10926 solver.cpp:245] Train net output #37: loss1/loss11 = 0.587435 (* 0.0272727 = 0.016021 loss) | |
| I0510 15:53:22.445569 10926 solver.cpp:245] Train net output #38: loss1/loss12 = 1.10881 (* 0.0272727 = 0.0302402 loss) | |
| I0510 15:53:22.445583 10926 solver.cpp:245] Train net output #39: loss1/loss13 = 0.659501 (* 0.0272727 = 0.0179864 loss) | |
| I0510 15:53:22.445618 10926 solver.cpp:245] Train net output #40: loss1/loss14 = 0.94307 (* 0.0272727 = 0.0257201 loss) | |
| I0510 15:53:22.445642 10926 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0262673 (* 0.0272727 = 0.00071638 loss) | |
| I0510 15:53:22.445657 10926 solver.cpp:245] Train net output #42: loss1/loss16 = 0.011912 (* 0.0272727 = 0.000324872 loss) | |
| I0510 15:53:22.445672 10926 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0116707 (* 0.0272727 = 0.000318291 loss) | |
| I0510 15:53:22.445686 10926 solver.cpp:245] Train net output #44: loss1/loss18 = 0.00960154 (* 0.0272727 = 0.00026186 loss) | |
| I0510 15:53:22.445701 10926 solver.cpp:245] Train net output #45: loss1/loss19 = 0.006809 (* 0.0272727 = 0.0001857 loss) | |
| I0510 15:53:22.445724 10926 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0062505 (* 0.0272727 = 0.000170468 loss) | |
| I0510 15:53:22.445739 10926 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00854171 (* 0.0272727 = 0.000232956 loss) | |
| I0510 15:53:22.445754 10926 solver.cpp:245] Train net output #48: loss1/loss22 = 0.00548318 (* 0.0272727 = 0.000149541 loss) | |
| I0510 15:53:22.445766 10926 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0188679 | |
| I0510 15:53:22.445778 10926 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0510 15:53:22.445791 10926 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0510 15:53:22.445801 10926 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0510 15:53:22.445813 10926 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0 | |
| I0510 15:53:22.445825 10926 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.25 | |
| I0510 15:53:22.445837 10926 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.375 | |
| I0510 15:53:22.445849 10926 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.625 | |
| I0510 15:53:22.445861 10926 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.75 | |
| I0510 15:53:22.445878 10926 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.875 | |
| I0510 15:53:22.445890 10926 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.875 | |
| I0510 15:53:22.445902 10926 solver.cpp:245] Train net output #60: loss2/accuracy11 = 0.875 | |
| I0510 15:53:22.445914 10926 solver.cpp:245] Train net output #61: loss2/accuracy12 = 0.875 | |
| I0510 15:53:22.445926 10926 solver.cpp:245] Train net output #62: loss2/accuracy13 = 0.875 | |
| I0510 15:53:22.445938 10926 solver.cpp:245] Train net output #63: loss2/accuracy14 = 0.875 | |
| I0510 15:53:22.445950 10926 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0510 15:53:22.445962 10926 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0510 15:53:22.445974 10926 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0510 15:53:22.445986 10926 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0510 15:53:22.445997 10926 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0510 15:53:22.446009 10926 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0510 15:53:22.446022 10926 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0510 15:53:22.446033 10926 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0510 15:53:22.446044 10926 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.698864 | |
| I0510 15:53:22.446058 10926 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.0943396 | |
| I0510 15:53:22.446074 10926 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 4.13904 (* 0.3 = 1.24171 loss) | |
| I0510 15:53:22.446089 10926 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.41687 (* 0.3 = 0.425062 loss) | |
| I0510 15:53:22.446102 10926 solver.cpp:245] Train net output #76: loss2/loss01 = 4.47578 (* 0.0272727 = 0.122067 loss) | |
| I0510 15:53:22.446116 10926 solver.cpp:245] Train net output #77: loss2/loss02 = 3.83662 (* 0.0272727 = 0.104635 loss) | |
| I0510 15:53:22.446142 10926 solver.cpp:245] Train net output #78: loss2/loss03 = 4.0143 (* 0.0272727 = 0.109481 loss) | |
| I0510 15:53:22.446158 10926 solver.cpp:245] Train net output #79: loss2/loss04 = 4.6236 (* 0.0272727 = 0.126098 loss) | |
| I0510 15:53:22.446172 10926 solver.cpp:245] Train net output #80: loss2/loss05 = 3.29294 (* 0.0272727 = 0.0898076 loss) | |
| I0510 15:53:22.446187 10926 solver.cpp:245] Train net output #81: loss2/loss06 = 2.84191 (* 0.0272727 = 0.0775065 loss) | |
| I0510 15:53:22.446200 10926 solver.cpp:245] Train net output #82: loss2/loss07 = 2.59605 (* 0.0272727 = 0.0708013 loss) | |
| I0510 15:53:22.446214 10926 solver.cpp:245] Train net output #83: loss2/loss08 = 1.7762 (* 0.0272727 = 0.0484417 loss) | |
| I0510 15:53:22.446228 10926 solver.cpp:245] Train net output #84: loss2/loss09 = 0.840191 (* 0.0272727 = 0.0229143 loss) | |
| I0510 15:53:22.446243 10926 solver.cpp:245] Train net output #85: loss2/loss10 = 1.01364 (* 0.0272727 = 0.0276448 loss) | |
| I0510 15:53:22.446256 10926 solver.cpp:245] Train net output #86: loss2/loss11 = 0.995344 (* 0.0272727 = 0.0271457 loss) | |
| I0510 15:53:22.446270 10926 solver.cpp:245] Train net output #87: loss2/loss12 = 0.781245 (* 0.0272727 = 0.0213067 loss) | |
| I0510 15:53:22.446285 10926 solver.cpp:245] Train net output #88: loss2/loss13 = 1.02297 (* 0.0272727 = 0.0278991 loss) | |
| I0510 15:53:22.446298 10926 solver.cpp:245] Train net output #89: loss2/loss14 = 1.13186 (* 0.0272727 = 0.0308689 loss) | |
| I0510 15:53:22.446312 10926 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0204043 (* 0.0272727 = 0.00055648 loss) | |
| I0510 15:53:22.446326 10926 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0209627 (* 0.0272727 = 0.000571711 loss) | |
| I0510 15:53:22.446341 10926 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0055809 (* 0.0272727 = 0.000152206 loss) | |
| I0510 15:53:22.446355 10926 solver.cpp:245] Train net output #93: loss2/loss18 = 0.00690015 (* 0.0272727 = 0.000188186 loss) | |
| I0510 15:53:22.446369 10926 solver.cpp:245] Train net output #94: loss2/loss19 = 0.00564448 (* 0.0272727 = 0.00015394 loss) | |
| I0510 15:53:22.446383 10926 solver.cpp:245] Train net output #95: loss2/loss20 = 0.0087592 (* 0.0272727 = 0.000238887 loss) | |
| I0510 15:53:22.446398 10926 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00653297 (* 0.0272727 = 0.000178172 loss) | |
| I0510 15:53:22.446413 10926 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00441089 (* 0.0272727 = 0.000120297 loss) | |
| I0510 15:53:22.446425 10926 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0188679 | |
| I0510 15:53:22.446439 10926 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0510 15:53:22.446450 10926 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.25 | |
| I0510 15:53:22.446461 10926 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0.125 | |
| I0510 15:53:22.446473 10926 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0 | |
| I0510 15:53:22.446485 10926 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.25 | |
| I0510 15:53:22.446497 10926 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.375 | |
| I0510 15:53:22.446506 10926 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.625 | |
| I0510 15:53:22.446514 10926 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.75 | |
| I0510 15:53:22.446527 10926 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.875 | |
| I0510 15:53:22.446539 10926 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.875 | |
| I0510 15:53:22.446550 10926 solver.cpp:245] Train net output #109: loss3/accuracy11 = 0.875 | |
| I0510 15:53:22.446563 10926 solver.cpp:245] Train net output #110: loss3/accuracy12 = 0.875 | |
| I0510 15:53:22.446573 10926 solver.cpp:245] Train net output #111: loss3/accuracy13 = 0.875 | |
| I0510 15:53:22.446585 10926 solver.cpp:245] Train net output #112: loss3/accuracy14 = 0.875 | |
| I0510 15:53:22.446596 10926 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0510 15:53:22.446609 10926 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0510 15:53:22.446630 10926 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0510 15:53:22.446650 10926 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0510 15:53:22.446660 10926 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0510 15:53:22.446672 10926 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0510 15:53:22.446683 10926 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0510 15:53:22.446696 10926 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0510 15:53:22.446712 10926 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.698864 | |
| I0510 15:53:22.446724 10926 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.113208 | |
| I0510 15:53:22.446738 10926 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 4.08945 (* 1 = 4.08945 loss) | |
| I0510 15:53:22.446753 10926 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.33935 (* 1 = 1.33935 loss) | |
| I0510 15:53:22.446765 10926 solver.cpp:245] Train net output #125: loss3/loss01 = 3.89584 (* 0.0909091 = 0.354167 loss) | |
| I0510 15:53:22.446779 10926 solver.cpp:245] Train net output #126: loss3/loss02 = 3.73682 (* 0.0909091 = 0.339711 loss) | |
| I0510 15:53:22.446794 10926 solver.cpp:245] Train net output #127: loss3/loss03 = 3.53234 (* 0.0909091 = 0.321122 loss) | |
| I0510 15:53:22.446807 10926 solver.cpp:245] Train net output #128: loss3/loss04 = 3.90153 (* 0.0909091 = 0.354685 loss) | |
| I0510 15:53:22.446821 10926 solver.cpp:245] Train net output #129: loss3/loss05 = 3.11937 (* 0.0909091 = 0.283579 loss) | |
| I0510 15:53:22.446835 10926 solver.cpp:245] Train net output #130: loss3/loss06 = 2.85559 (* 0.0909091 = 0.259599 loss) | |
| I0510 15:53:22.446849 10926 solver.cpp:245] Train net output #131: loss3/loss07 = 2.0519 (* 0.0909091 = 0.186536 loss) | |
| I0510 15:53:22.446862 10926 solver.cpp:245] Train net output #132: loss3/loss08 = 1.65337 (* 0.0909091 = 0.150306 loss) | |
| I0510 15:53:22.446877 10926 solver.cpp:245] Train net output #133: loss3/loss09 = 0.798769 (* 0.0909091 = 0.0726154 loss) | |
| I0510 15:53:22.446890 10926 solver.cpp:245] Train net output #134: loss3/loss10 = 1.03239 (* 0.0909091 = 0.0938537 loss) | |
| I0510 15:53:22.446904 10926 solver.cpp:245] Train net output #135: loss3/loss11 = 0.983643 (* 0.0909091 = 0.0894221 loss) | |
| I0510 15:53:22.446918 10926 solver.cpp:245] Train net output #136: loss3/loss12 = 1.11963 (* 0.0909091 = 0.101785 loss) | |
| I0510 15:53:22.446935 10926 solver.cpp:245] Train net output #137: loss3/loss13 = 1.18875 (* 0.0909091 = 0.108068 loss) | |
| I0510 15:53:22.446950 10926 solver.cpp:245] Train net output #138: loss3/loss14 = 1.29963 (* 0.0909091 = 0.118148 loss) | |
| I0510 15:53:22.446964 10926 solver.cpp:245] Train net output #139: loss3/loss15 = 0.0124572 (* 0.0909091 = 0.00113248 loss) | |
| I0510 15:53:22.446979 10926 solver.cpp:245] Train net output #140: loss3/loss16 = 0.00407605 (* 0.0909091 = 0.00037055 loss) | |
| I0510 15:53:22.446992 10926 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00327052 (* 0.0909091 = 0.00029732 loss) | |
| I0510 15:53:22.447006 10926 solver.cpp:245] Train net output #142: loss3/loss18 = 0.0025971 (* 0.0909091 = 0.0002361 loss) | |
| I0510 15:53:22.447021 10926 solver.cpp:245] Train net output #143: loss3/loss19 = 0.000861217 (* 0.0909091 = 7.82925e-05 loss) | |
| I0510 15:53:22.447036 10926 solver.cpp:245] Train net output #144: loss3/loss20 = 0.000896771 (* 0.0909091 = 8.15246e-05 loss) | |
| I0510 15:53:22.447049 10926 solver.cpp:245] Train net output #145: loss3/loss21 = 0.0012253 (* 0.0909091 = 0.000111391 loss) | |
| I0510 15:53:22.447064 10926 solver.cpp:245] Train net output #146: loss3/loss22 = 0.000944721 (* 0.0909091 = 8.58838e-05 loss) | |
| I0510 15:53:22.447077 10926 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0510 15:53:22.447088 10926 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0510 15:53:22.447099 10926 solver.cpp:245] Train net output #149: total_confidence = 1.58273e-06 | |
| I0510 15:53:22.447124 10926 solver.cpp:245] Train net output #150: total_confidence_not_rec = 6.90339e-05 | |
| I0510 15:53:22.447140 10926 sgd_solver.cpp:106] Iteration 5500, lr = 0.001 | |
| I0510 15:53:26.075044 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 31.8304 > 30) by scale factor 0.942496 | |
| I0510 15:53:58.257207 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 43.5143 > 30) by scale factor 0.689429 | |
| I0510 15:54:01.491592 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 31.0576 > 30) by scale factor 0.965948 | |
| I0510 15:55:37.366696 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 31.2997 > 30) by scale factor 0.958476 | |
| I0510 15:55:49.632709 10926 solver.cpp:229] Iteration 6000, loss = 11.4974 | |
| I0510 15:55:49.632781 10926 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0540541 | |
| I0510 15:55:49.632800 10926 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.25 | |
| I0510 15:55:49.632814 10926 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0510 15:55:49.632827 10926 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0510 15:55:49.632839 10926 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.25 | |
| I0510 15:55:49.632851 10926 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.5 | |
| I0510 15:55:49.632864 10926 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.625 | |
| I0510 15:55:49.632876 10926 solver.cpp:245] Train net output #7: loss1/accuracy07 = 1 | |
| I0510 15:55:49.632889 10926 solver.cpp:245] Train net output #8: loss1/accuracy08 = 1 | |
| I0510 15:55:49.632900 10926 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0510 15:55:49.632912 10926 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0510 15:55:49.632925 10926 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0510 15:55:49.632936 10926 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0510 15:55:49.632948 10926 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0510 15:55:49.632961 10926 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0510 15:55:49.632973 10926 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0510 15:55:49.632984 10926 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0510 15:55:49.632997 10926 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0510 15:55:49.633008 10926 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0510 15:55:49.633020 10926 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0510 15:55:49.633033 10926 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0510 15:55:49.633044 10926 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0510 15:55:49.633055 10926 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0510 15:55:49.633069 10926 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.795455 | |
| I0510 15:55:49.633080 10926 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.27027 | |
| I0510 15:55:49.633096 10926 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.57156 (* 0.3 = 1.07147 loss) | |
| I0510 15:55:49.633111 10926 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 0.896642 (* 0.3 = 0.268993 loss) | |
| I0510 15:55:49.633142 10926 solver.cpp:245] Train net output #27: loss1/loss01 = 3.18619 (* 0.0272727 = 0.086896 loss) | |
| I0510 15:55:49.633158 10926 solver.cpp:245] Train net output #28: loss1/loss02 = 3.90708 (* 0.0272727 = 0.106557 loss) | |
| I0510 15:55:49.633173 10926 solver.cpp:245] Train net output #29: loss1/loss03 = 3.44659 (* 0.0272727 = 0.0939978 loss) | |
| I0510 15:55:49.633188 10926 solver.cpp:245] Train net output #30: loss1/loss04 = 3.04188 (* 0.0272727 = 0.0829602 loss) | |
| I0510 15:55:49.633203 10926 solver.cpp:245] Train net output #31: loss1/loss05 = 2.18334 (* 0.0272727 = 0.0595455 loss) | |
| I0510 15:55:49.633216 10926 solver.cpp:245] Train net output #32: loss1/loss06 = 1.94492 (* 0.0272727 = 0.0530432 loss) | |
| I0510 15:55:49.633230 10926 solver.cpp:245] Train net output #33: loss1/loss07 = 0.385618 (* 0.0272727 = 0.0105169 loss) | |
| I0510 15:55:49.633245 10926 solver.cpp:245] Train net output #34: loss1/loss08 = 0.177178 (* 0.0272727 = 0.00483212 loss) | |
| I0510 15:55:49.633260 10926 solver.cpp:245] Train net output #35: loss1/loss09 = 0.0789239 (* 0.0272727 = 0.00215247 loss) | |
| I0510 15:55:49.633275 10926 solver.cpp:245] Train net output #36: loss1/loss10 = 0.0500561 (* 0.0272727 = 0.00136517 loss) | |
| I0510 15:55:49.633288 10926 solver.cpp:245] Train net output #37: loss1/loss11 = 0.0634303 (* 0.0272727 = 0.00172992 loss) | |
| I0510 15:55:49.633303 10926 solver.cpp:245] Train net output #38: loss1/loss12 = 0.043179 (* 0.0272727 = 0.00117761 loss) | |
| I0510 15:55:49.633357 10926 solver.cpp:245] Train net output #39: loss1/loss13 = 0.029544 (* 0.0272727 = 0.000805744 loss) | |
| I0510 15:55:49.633373 10926 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0363348 (* 0.0272727 = 0.000990949 loss) | |
| I0510 15:55:49.633388 10926 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0348944 (* 0.0272727 = 0.000951667 loss) | |
| I0510 15:55:49.633402 10926 solver.cpp:245] Train net output #42: loss1/loss16 = 0.015367 (* 0.0272727 = 0.000419101 loss) | |
| I0510 15:55:49.633417 10926 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0126962 (* 0.0272727 = 0.000346261 loss) | |
| I0510 15:55:49.633431 10926 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0175998 (* 0.0272727 = 0.000479994 loss) | |
| I0510 15:55:49.633445 10926 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0101643 (* 0.0272727 = 0.000277208 loss) | |
| I0510 15:55:49.633460 10926 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00951579 (* 0.0272727 = 0.000259521 loss) | |
| I0510 15:55:49.633474 10926 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00729933 (* 0.0272727 = 0.000199073 loss) | |
| I0510 15:55:49.633489 10926 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0101206 (* 0.0272727 = 0.000276016 loss) | |
| I0510 15:55:49.633502 10926 solver.cpp:245] Train net output #49: loss2/accuracy = 0.027027 | |
| I0510 15:55:49.633514 10926 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.25 | |
| I0510 15:55:49.633527 10926 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0.125 | |
| I0510 15:55:49.633539 10926 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0.125 | |
| I0510 15:55:49.633548 10926 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.25 | |
| I0510 15:55:49.633555 10926 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.5 | |
| I0510 15:55:49.633569 10926 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.625 | |
| I0510 15:55:49.633581 10926 solver.cpp:245] Train net output #56: loss2/accuracy07 = 1 | |
| I0510 15:55:49.633594 10926 solver.cpp:245] Train net output #57: loss2/accuracy08 = 1 | |
| I0510 15:55:49.633605 10926 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0510 15:55:49.633616 10926 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0510 15:55:49.633628 10926 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0510 15:55:49.633641 10926 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0510 15:55:49.633652 10926 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0510 15:55:49.633664 10926 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0510 15:55:49.633678 10926 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0510 15:55:49.633692 10926 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0510 15:55:49.633703 10926 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0510 15:55:49.633714 10926 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0510 15:55:49.633726 10926 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0510 15:55:49.633738 10926 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0510 15:55:49.633750 10926 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0510 15:55:49.633761 10926 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0510 15:55:49.633774 10926 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.789773 | |
| I0510 15:55:49.633785 10926 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.216216 | |
| I0510 15:55:49.633800 10926 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.55323 (* 0.3 = 1.06597 loss) | |
| I0510 15:55:49.633813 10926 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 0.958208 (* 0.3 = 0.287463 loss) | |
| I0510 15:55:49.633828 10926 solver.cpp:245] Train net output #76: loss2/loss01 = 3.14434 (* 0.0272727 = 0.0857547 loss) | |
| I0510 15:55:49.633842 10926 solver.cpp:245] Train net output #77: loss2/loss02 = 3.35517 (* 0.0272727 = 0.0915045 loss) | |
| I0510 15:55:49.633868 10926 solver.cpp:245] Train net output #78: loss2/loss03 = 3.69137 (* 0.0272727 = 0.100674 loss) | |
| I0510 15:55:49.633883 10926 solver.cpp:245] Train net output #79: loss2/loss04 = 3.4492 (* 0.0272727 = 0.0940691 loss) | |
| I0510 15:55:49.633898 10926 solver.cpp:245] Train net output #80: loss2/loss05 = 2.51219 (* 0.0272727 = 0.0685143 loss) | |
| I0510 15:55:49.633913 10926 solver.cpp:245] Train net output #81: loss2/loss06 = 1.96492 (* 0.0272727 = 0.0535886 loss) | |
| I0510 15:55:49.633926 10926 solver.cpp:245] Train net output #82: loss2/loss07 = 0.430924 (* 0.0272727 = 0.0117525 loss) | |
| I0510 15:55:49.633941 10926 solver.cpp:245] Train net output #83: loss2/loss08 = 0.148592 (* 0.0272727 = 0.00405251 loss) | |
| I0510 15:55:49.633955 10926 solver.cpp:245] Train net output #84: loss2/loss09 = 0.0709304 (* 0.0272727 = 0.00193446 loss) | |
| I0510 15:55:49.633970 10926 solver.cpp:245] Train net output #85: loss2/loss10 = 0.0901023 (* 0.0272727 = 0.00245733 loss) | |
| I0510 15:55:49.633985 10926 solver.cpp:245] Train net output #86: loss2/loss11 = 0.0780913 (* 0.0272727 = 0.00212976 loss) | |
| I0510 15:55:49.633998 10926 solver.cpp:245] Train net output #87: loss2/loss12 = 0.0575612 (* 0.0272727 = 0.00156985 loss) | |
| I0510 15:55:49.634013 10926 solver.cpp:245] Train net output #88: loss2/loss13 = 0.0580801 (* 0.0272727 = 0.001584 loss) | |
| I0510 15:55:49.634027 10926 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0314329 (* 0.0272727 = 0.00085726 loss) | |
| I0510 15:55:49.634042 10926 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0551219 (* 0.0272727 = 0.00150332 loss) | |
| I0510 15:55:49.634057 10926 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0149412 (* 0.0272727 = 0.000407486 loss) | |
| I0510 15:55:49.634071 10926 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0120994 (* 0.0272727 = 0.000329985 loss) | |
| I0510 15:55:49.634085 10926 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0120556 (* 0.0272727 = 0.000328788 loss) | |
| I0510 15:55:49.634100 10926 solver.cpp:245] Train net output #94: loss2/loss19 = 0.00747257 (* 0.0272727 = 0.000203797 loss) | |
| I0510 15:55:49.634114 10926 solver.cpp:245] Train net output #95: loss2/loss20 = 0.0114088 (* 0.0272727 = 0.000311149 loss) | |
| I0510 15:55:49.634129 10926 solver.cpp:245] Train net output #96: loss2/loss21 = 0.0144904 (* 0.0272727 = 0.000395192 loss) | |
| I0510 15:55:49.634143 10926 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00963252 (* 0.0272727 = 0.000262705 loss) | |
| I0510 15:55:49.634156 10926 solver.cpp:245] Train net output #98: loss3/accuracy = 0.135135 | |
| I0510 15:55:49.634169 10926 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.375 | |
| I0510 15:55:49.634186 10926 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.125 | |
| I0510 15:55:49.634197 10926 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0510 15:55:49.634210 10926 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.375 | |
| I0510 15:55:49.634222 10926 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.625 | |
| I0510 15:55:49.634234 10926 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.625 | |
| I0510 15:55:49.634246 10926 solver.cpp:245] Train net output #105: loss3/accuracy07 = 1 | |
| I0510 15:55:49.634258 10926 solver.cpp:245] Train net output #106: loss3/accuracy08 = 1 | |
| I0510 15:55:49.634270 10926 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0510 15:55:49.634282 10926 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0510 15:55:49.634294 10926 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0510 15:55:49.634306 10926 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0510 15:55:49.634317 10926 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0510 15:55:49.634330 10926 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0510 15:55:49.634341 10926 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0510 15:55:49.634363 10926 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0510 15:55:49.634377 10926 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0510 15:55:49.634389 10926 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0510 15:55:49.634402 10926 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0510 15:55:49.634413 10926 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0510 15:55:49.634424 10926 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0510 15:55:49.634436 10926 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0510 15:55:49.634449 10926 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.8125 | |
| I0510 15:55:49.634460 10926 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.216216 | |
| I0510 15:55:49.634474 10926 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.3458 (* 1 = 3.3458 loss) | |
| I0510 15:55:49.634488 10926 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 0.862552 (* 1 = 0.862552 loss) | |
| I0510 15:55:49.634503 10926 solver.cpp:245] Train net output #125: loss3/loss01 = 2.98485 (* 0.0909091 = 0.27135 loss) | |
| I0510 15:55:49.634517 10926 solver.cpp:245] Train net output #126: loss3/loss02 = 3.18735 (* 0.0909091 = 0.289759 loss) | |
| I0510 15:55:49.634531 10926 solver.cpp:245] Train net output #127: loss3/loss03 = 3.34888 (* 0.0909091 = 0.304444 loss) | |
| I0510 15:55:49.634546 10926 solver.cpp:245] Train net output #128: loss3/loss04 = 2.76153 (* 0.0909091 = 0.251048 loss) | |
| I0510 15:55:49.634559 10926 solver.cpp:245] Train net output #129: loss3/loss05 = 2.13954 (* 0.0909091 = 0.194503 loss) | |
| I0510 15:55:49.634573 10926 solver.cpp:245] Train net output #130: loss3/loss06 = 2.01236 (* 0.0909091 = 0.182942 loss) | |
| I0510 15:55:49.634588 10926 solver.cpp:245] Train net output #131: loss3/loss07 = 0.287189 (* 0.0909091 = 0.0261081 loss) | |
| I0510 15:55:49.634601 10926 solver.cpp:245] Train net output #132: loss3/loss08 = 0.123786 (* 0.0909091 = 0.0112532 loss) | |
| I0510 15:55:49.634616 10926 solver.cpp:245] Train net output #133: loss3/loss09 = 0.0376713 (* 0.0909091 = 0.00342467 loss) | |
| I0510 15:55:49.634631 10926 solver.cpp:245] Train net output #134: loss3/loss10 = 0.0271788 (* 0.0909091 = 0.0024708 loss) | |
| I0510 15:55:49.634645 10926 solver.cpp:245] Train net output #135: loss3/loss11 = 0.0165833 (* 0.0909091 = 0.00150758 loss) | |
| I0510 15:55:49.634660 10926 solver.cpp:245] Train net output #136: loss3/loss12 = 0.0218378 (* 0.0909091 = 0.00198525 loss) | |
| I0510 15:55:49.634675 10926 solver.cpp:245] Train net output #137: loss3/loss13 = 0.0109829 (* 0.0909091 = 0.000998446 loss) | |
| I0510 15:55:49.634688 10926 solver.cpp:245] Train net output #138: loss3/loss14 = 0.00906823 (* 0.0909091 = 0.000824385 loss) | |
| I0510 15:55:49.634702 10926 solver.cpp:245] Train net output #139: loss3/loss15 = 0.0074489 (* 0.0909091 = 0.000677172 loss) | |
| I0510 15:55:49.634716 10926 solver.cpp:245] Train net output #140: loss3/loss16 = 0.00396115 (* 0.0909091 = 0.000360104 loss) | |
| I0510 15:55:49.634734 10926 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00176735 (* 0.0909091 = 0.000160668 loss) | |
| I0510 15:55:49.634748 10926 solver.cpp:245] Train net output #142: loss3/loss18 = 0.0013417 (* 0.0909091 = 0.000121972 loss) | |
| I0510 15:55:49.634763 10926 solver.cpp:245] Train net output #143: loss3/loss19 = 0.000800598 (* 0.0909091 = 7.27816e-05 loss) | |
| I0510 15:55:49.634778 10926 solver.cpp:245] Train net output #144: loss3/loss20 = 0.000511416 (* 0.0909091 = 4.64924e-05 loss) | |
| I0510 15:55:49.634793 10926 solver.cpp:245] Train net output #145: loss3/loss21 = 0.000395169 (* 0.0909091 = 3.59244e-05 loss) | |
| I0510 15:55:49.634806 10926 solver.cpp:245] Train net output #146: loss3/loss22 = 0.000456314 (* 0.0909091 = 4.14831e-05 loss) | |
| I0510 15:55:49.634819 10926 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0510 15:55:49.634831 10926 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0510 15:55:49.634852 10926 solver.cpp:245] Train net output #149: total_confidence = 7.50698e-05 | |
| I0510 15:55:49.634866 10926 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.000344607 | |
| I0510 15:55:49.634881 10926 sgd_solver.cpp:106] Iteration 6000, lr = 0.001 | |
| I0510 15:55:55.905179 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 44.1151 > 30) by scale factor 0.680039 | |
| I0510 15:56:36.135340 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 30.0357 > 30) by scale factor 0.998811 | |
| I0510 15:56:54.757293 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 38.2064 > 30) by scale factor 0.785209 | |
| I0510 15:57:06.600179 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 30.0847 > 30) by scale factor 0.997186 | |
| I0510 15:57:15.191298 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 57.5695 > 30) by scale factor 0.521109 | |
| I0510 15:57:16.083802 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 39.9031 > 30) by scale factor 0.751822 | |
| I0510 15:57:41.452366 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 33.7679 > 30) by scale factor 0.888417 | |
| I0510 15:57:56.248307 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 47.4067 > 30) by scale factor 0.632822 | |
| I0510 15:58:17.141013 10926 solver.cpp:229] Iteration 6500, loss = 11.3125 | |
| I0510 15:58:17.141140 10926 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0714286 | |
| I0510 15:58:17.141161 10926 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.125 | |
| I0510 15:58:17.141175 10926 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0510 15:58:17.141187 10926 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.125 | |
| I0510 15:58:17.141199 10926 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.25 | |
| I0510 15:58:17.141211 10926 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.25 | |
| I0510 15:58:17.141223 10926 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.375 | |
| I0510 15:58:17.141235 10926 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.5 | |
| I0510 15:58:17.141247 10926 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.75 | |
| I0510 15:58:17.141259 10926 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.75 | |
| I0510 15:58:17.141271 10926 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.875 | |
| I0510 15:58:17.141284 10926 solver.cpp:245] Train net output #11: loss1/accuracy11 = 0.875 | |
| I0510 15:58:17.141295 10926 solver.cpp:245] Train net output #12: loss1/accuracy12 = 0.875 | |
| I0510 15:58:17.141307 10926 solver.cpp:245] Train net output #13: loss1/accuracy13 = 0.875 | |
| I0510 15:58:17.141320 10926 solver.cpp:245] Train net output #14: loss1/accuracy14 = 0.875 | |
| I0510 15:58:17.141332 10926 solver.cpp:245] Train net output #15: loss1/accuracy15 = 0.875 | |
| I0510 15:58:17.141345 10926 solver.cpp:245] Train net output #16: loss1/accuracy16 = 0.875 | |
| I0510 15:58:17.141356 10926 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0510 15:58:17.141368 10926 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0510 15:58:17.141381 10926 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0510 15:58:17.141392 10926 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0510 15:58:17.141404 10926 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0510 15:58:17.141415 10926 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0510 15:58:17.141427 10926 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.6875 | |
| I0510 15:58:17.141439 10926 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.25 | |
| I0510 15:58:17.141455 10926 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.6935 (* 0.3 = 1.10805 loss) | |
| I0510 15:58:17.141470 10926 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.34395 (* 0.3 = 0.403184 loss) | |
| I0510 15:58:17.141485 10926 solver.cpp:245] Train net output #27: loss1/loss01 = 3.40583 (* 0.0272727 = 0.0928862 loss) | |
| I0510 15:58:17.141499 10926 solver.cpp:245] Train net output #28: loss1/loss02 = 3.62245 (* 0.0272727 = 0.0987941 loss) | |
| I0510 15:58:17.141520 10926 solver.cpp:245] Train net output #29: loss1/loss03 = 3.71634 (* 0.0272727 = 0.101355 loss) | |
| I0510 15:58:17.141533 10926 solver.cpp:245] Train net output #30: loss1/loss04 = 3.68711 (* 0.0272727 = 0.100558 loss) | |
| I0510 15:58:17.141547 10926 solver.cpp:245] Train net output #31: loss1/loss05 = 2.61323 (* 0.0272727 = 0.07127 loss) | |
| I0510 15:58:17.141561 10926 solver.cpp:245] Train net output #32: loss1/loss06 = 3.00277 (* 0.0272727 = 0.0818937 loss) | |
| I0510 15:58:17.141582 10926 solver.cpp:245] Train net output #33: loss1/loss07 = 2.5165 (* 0.0272727 = 0.0686317 loss) | |
| I0510 15:58:17.141595 10926 solver.cpp:245] Train net output #34: loss1/loss08 = 1.156 (* 0.0272727 = 0.0315274 loss) | |
| I0510 15:58:17.141609 10926 solver.cpp:245] Train net output #35: loss1/loss09 = 1.00156 (* 0.0272727 = 0.0273152 loss) | |
| I0510 15:58:17.141623 10926 solver.cpp:245] Train net output #36: loss1/loss10 = 0.786286 (* 0.0272727 = 0.0214442 loss) | |
| I0510 15:58:17.141646 10926 solver.cpp:245] Train net output #37: loss1/loss11 = 0.663032 (* 0.0272727 = 0.0180827 loss) | |
| I0510 15:58:17.141660 10926 solver.cpp:245] Train net output #38: loss1/loss12 = 0.746036 (* 0.0272727 = 0.0203464 loss) | |
| I0510 15:58:17.141674 10926 solver.cpp:245] Train net output #39: loss1/loss13 = 0.824884 (* 0.0272727 = 0.0224968 loss) | |
| I0510 15:58:17.141707 10926 solver.cpp:245] Train net output #40: loss1/loss14 = 0.899703 (* 0.0272727 = 0.0245374 loss) | |
| I0510 15:58:17.141727 10926 solver.cpp:245] Train net output #41: loss1/loss15 = 0.785897 (* 0.0272727 = 0.0214336 loss) | |
| I0510 15:58:17.141741 10926 solver.cpp:245] Train net output #42: loss1/loss16 = 0.943019 (* 0.0272727 = 0.0257187 loss) | |
| I0510 15:58:17.141757 10926 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0246854 (* 0.0272727 = 0.000673237 loss) | |
| I0510 15:58:17.141770 10926 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0122044 (* 0.0272727 = 0.000332846 loss) | |
| I0510 15:58:17.141784 10926 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0142658 (* 0.0272727 = 0.000389068 loss) | |
| I0510 15:58:17.141798 10926 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0269493 (* 0.0272727 = 0.00073498 loss) | |
| I0510 15:58:17.141813 10926 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0178461 (* 0.0272727 = 0.000486713 loss) | |
| I0510 15:58:17.141826 10926 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0237771 (* 0.0272727 = 0.000648466 loss) | |
| I0510 15:58:17.141839 10926 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0892857 | |
| I0510 15:58:17.141851 10926 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0510 15:58:17.141863 10926 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0510 15:58:17.141880 10926 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0510 15:58:17.141891 10926 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.125 | |
| I0510 15:58:17.141903 10926 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.375 | |
| I0510 15:58:17.141916 10926 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.375 | |
| I0510 15:58:17.141927 10926 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.5 | |
| I0510 15:58:17.141938 10926 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.75 | |
| I0510 15:58:17.141950 10926 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.75 | |
| I0510 15:58:17.141962 10926 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.875 | |
| I0510 15:58:17.141973 10926 solver.cpp:245] Train net output #60: loss2/accuracy11 = 0.875 | |
| I0510 15:58:17.141985 10926 solver.cpp:245] Train net output #61: loss2/accuracy12 = 0.875 | |
| I0510 15:58:17.141998 10926 solver.cpp:245] Train net output #62: loss2/accuracy13 = 0.875 | |
| I0510 15:58:17.142009 10926 solver.cpp:245] Train net output #63: loss2/accuracy14 = 0.875 | |
| I0510 15:58:17.142021 10926 solver.cpp:245] Train net output #64: loss2/accuracy15 = 0.875 | |
| I0510 15:58:17.142033 10926 solver.cpp:245] Train net output #65: loss2/accuracy16 = 0.875 | |
| I0510 15:58:17.142045 10926 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0510 15:58:17.142056 10926 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0510 15:58:17.142067 10926 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0510 15:58:17.142079 10926 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0510 15:58:17.142091 10926 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0510 15:58:17.142102 10926 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0510 15:58:17.142113 10926 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.693182 | |
| I0510 15:58:17.142125 10926 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.125 | |
| I0510 15:58:17.142139 10926 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.49491 (* 0.3 = 1.04847 loss) | |
| I0510 15:58:17.142156 10926 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.35274 (* 0.3 = 0.405821 loss) | |
| I0510 15:58:17.142171 10926 solver.cpp:245] Train net output #76: loss2/loss01 = 3.41254 (* 0.0272727 = 0.0930694 loss) | |
| I0510 15:58:17.142186 10926 solver.cpp:245] Train net output #77: loss2/loss02 = 3.6408 (* 0.0272727 = 0.0992945 loss) | |
| I0510 15:58:17.142212 10926 solver.cpp:245] Train net output #78: loss2/loss03 = 3.40101 (* 0.0272727 = 0.0927548 loss) | |
| I0510 15:58:17.142228 10926 solver.cpp:245] Train net output #79: loss2/loss04 = 3.56555 (* 0.0272727 = 0.0972422 loss) | |
| I0510 15:58:17.142243 10926 solver.cpp:245] Train net output #80: loss2/loss05 = 2.6745 (* 0.0272727 = 0.0729409 loss) | |
| I0510 15:58:17.142257 10926 solver.cpp:245] Train net output #81: loss2/loss06 = 2.85359 (* 0.0272727 = 0.0778251 loss) | |
| I0510 15:58:17.142277 10926 solver.cpp:245] Train net output #82: loss2/loss07 = 2.32291 (* 0.0272727 = 0.0633521 loss) | |
| I0510 15:58:17.142292 10926 solver.cpp:245] Train net output #83: loss2/loss08 = 1.05787 (* 0.0272727 = 0.0288511 loss) | |
| I0510 15:58:17.142305 10926 solver.cpp:245] Train net output #84: loss2/loss09 = 0.874709 (* 0.0272727 = 0.0238557 loss) | |
| I0510 15:58:17.142319 10926 solver.cpp:245] Train net output #85: loss2/loss10 = 0.681093 (* 0.0272727 = 0.0185753 loss) | |
| I0510 15:58:17.142335 10926 solver.cpp:245] Train net output #86: loss2/loss11 = 0.706761 (* 0.0272727 = 0.0192753 loss) | |
| I0510 15:58:17.142350 10926 solver.cpp:245] Train net output #87: loss2/loss12 = 0.615263 (* 0.0272727 = 0.0167799 loss) | |
| I0510 15:58:17.142364 10926 solver.cpp:245] Train net output #88: loss2/loss13 = 0.898467 (* 0.0272727 = 0.0245036 loss) | |
| I0510 15:58:17.142379 10926 solver.cpp:245] Train net output #89: loss2/loss14 = 0.595959 (* 0.0272727 = 0.0162534 loss) | |
| I0510 15:58:17.142392 10926 solver.cpp:245] Train net output #90: loss2/loss15 = 0.768253 (* 0.0272727 = 0.0209523 loss) | |
| I0510 15:58:17.142406 10926 solver.cpp:245] Train net output #91: loss2/loss16 = 0.943258 (* 0.0272727 = 0.0257252 loss) | |
| I0510 15:58:17.142419 10926 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0147672 (* 0.0272727 = 0.000402741 loss) | |
| I0510 15:58:17.142433 10926 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0175093 (* 0.0272727 = 0.000477528 loss) | |
| I0510 15:58:17.142447 10926 solver.cpp:245] Train net output #94: loss2/loss19 = 0.012088 (* 0.0272727 = 0.000329673 loss) | |
| I0510 15:58:17.142462 10926 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00744986 (* 0.0272727 = 0.000203178 loss) | |
| I0510 15:58:17.142475 10926 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00625811 (* 0.0272727 = 0.000170676 loss) | |
| I0510 15:58:17.142489 10926 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00572506 (* 0.0272727 = 0.000156138 loss) | |
| I0510 15:58:17.142501 10926 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0714286 | |
| I0510 15:58:17.142513 10926 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.375 | |
| I0510 15:58:17.142526 10926 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0510 15:58:17.142537 10926 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0.125 | |
| I0510 15:58:17.142549 10926 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.25 | |
| I0510 15:58:17.142561 10926 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.375 | |
| I0510 15:58:17.142573 10926 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.375 | |
| I0510 15:58:17.142585 10926 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.5 | |
| I0510 15:58:17.142596 10926 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.75 | |
| I0510 15:58:17.142607 10926 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.75 | |
| I0510 15:58:17.142619 10926 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.875 | |
| I0510 15:58:17.142630 10926 solver.cpp:245] Train net output #109: loss3/accuracy11 = 0.875 | |
| I0510 15:58:17.142648 10926 solver.cpp:245] Train net output #110: loss3/accuracy12 = 0.875 | |
| I0510 15:58:17.142659 10926 solver.cpp:245] Train net output #111: loss3/accuracy13 = 0.875 | |
| I0510 15:58:17.142671 10926 solver.cpp:245] Train net output #112: loss3/accuracy14 = 0.875 | |
| I0510 15:58:17.142683 10926 solver.cpp:245] Train net output #113: loss3/accuracy15 = 0.875 | |
| I0510 15:58:17.142693 10926 solver.cpp:245] Train net output #114: loss3/accuracy16 = 0.875 | |
| I0510 15:58:17.142715 10926 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0510 15:58:17.142729 10926 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0510 15:58:17.142740 10926 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0510 15:58:17.142752 10926 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0510 15:58:17.142763 10926 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0510 15:58:17.142776 10926 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0510 15:58:17.142786 10926 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.698864 | |
| I0510 15:58:17.142798 10926 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.25 | |
| I0510 15:58:17.142812 10926 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.28718 (* 1 = 3.28718 loss) | |
| I0510 15:58:17.142825 10926 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.18761 (* 1 = 1.18761 loss) | |
| I0510 15:58:17.142839 10926 solver.cpp:245] Train net output #125: loss3/loss01 = 2.77586 (* 0.0909091 = 0.252351 loss) | |
| I0510 15:58:17.142853 10926 solver.cpp:245] Train net output #126: loss3/loss02 = 3.08439 (* 0.0909091 = 0.280399 loss) | |
| I0510 15:58:17.142868 10926 solver.cpp:245] Train net output #127: loss3/loss03 = 3.41137 (* 0.0909091 = 0.310125 loss) | |
| I0510 15:58:17.142881 10926 solver.cpp:245] Train net output #128: loss3/loss04 = 3.32279 (* 0.0909091 = 0.302072 loss) | |
| I0510 15:58:17.142895 10926 solver.cpp:245] Train net output #129: loss3/loss05 = 2.49672 (* 0.0909091 = 0.226974 loss) | |
| I0510 15:58:17.142910 10926 solver.cpp:245] Train net output #130: loss3/loss06 = 2.64704 (* 0.0909091 = 0.24064 loss) | |
| I0510 15:58:17.142922 10926 solver.cpp:245] Train net output #131: loss3/loss07 = 2.10218 (* 0.0909091 = 0.191107 loss) | |
| I0510 15:58:17.142940 10926 solver.cpp:245] Train net output #132: loss3/loss08 = 1.01232 (* 0.0909091 = 0.0920292 loss) | |
| I0510 15:58:17.142956 10926 solver.cpp:245] Train net output #133: loss3/loss09 = 0.694821 (* 0.0909091 = 0.0631656 loss) | |
| I0510 15:58:17.142969 10926 solver.cpp:245] Train net output #134: loss3/loss10 = 0.742231 (* 0.0909091 = 0.0674755 loss) | |
| I0510 15:58:17.142983 10926 solver.cpp:245] Train net output #135: loss3/loss11 = 0.460695 (* 0.0909091 = 0.0418814 loss) | |
| I0510 15:58:17.142997 10926 solver.cpp:245] Train net output #136: loss3/loss12 = 0.749771 (* 0.0909091 = 0.068161 loss) | |
| I0510 15:58:17.143012 10926 solver.cpp:245] Train net output #137: loss3/loss13 = 0.543435 (* 0.0909091 = 0.0494032 loss) | |
| I0510 15:58:17.143031 10926 solver.cpp:245] Train net output #138: loss3/loss14 = 0.466269 (* 0.0909091 = 0.0423881 loss) | |
| I0510 15:58:17.143044 10926 solver.cpp:245] Train net output #139: loss3/loss15 = 0.577343 (* 0.0909091 = 0.0524858 loss) | |
| I0510 15:58:17.143059 10926 solver.cpp:245] Train net output #140: loss3/loss16 = 0.78267 (* 0.0909091 = 0.0711518 loss) | |
| I0510 15:58:17.143072 10926 solver.cpp:245] Train net output #141: loss3/loss17 = 0.0108093 (* 0.0909091 = 0.000982664 loss) | |
| I0510 15:58:17.143093 10926 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00631035 (* 0.0909091 = 0.000573668 loss) | |
| I0510 15:58:17.143107 10926 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00275926 (* 0.0909091 = 0.000250842 loss) | |
| I0510 15:58:17.143122 10926 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00245022 (* 0.0909091 = 0.000222747 loss) | |
| I0510 15:58:17.143136 10926 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00147856 (* 0.0909091 = 0.000134414 loss) | |
| I0510 15:58:17.143151 10926 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00185476 (* 0.0909091 = 0.000168615 loss) | |
| I0510 15:58:17.143163 10926 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0510 15:58:17.143175 10926 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0510 15:58:17.143187 10926 solver.cpp:245] Train net output #149: total_confidence = 0.000175652 | |
| I0510 15:58:17.143213 10926 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.00289508 | |
| I0510 15:58:17.143229 10926 sgd_solver.cpp:106] Iteration 6500, lr = 0.001 | |
| I0510 15:59:37.306625 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 36.9452 > 30) by scale factor 0.812015 | |
| I0510 15:59:37.609374 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 50.5223 > 30) by scale factor 0.593797 | |
| I0510 15:59:37.905374 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 32.0701 > 30) by scale factor 0.935452 | |
| I0510 16:00:05.355808 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 31.9624 > 30) by scale factor 0.938602 | |
| I0510 16:00:09.463171 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 41.6752 > 30) by scale factor 0.719853 | |
| I0510 16:00:26.529497 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 38.5614 > 30) by scale factor 0.777981 | |
| I0510 16:00:30.355643 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 31.73 > 30) by scale factor 0.945478 | |
| I0510 16:00:44.690884 10926 solver.cpp:229] Iteration 7000, loss = 11.1378 | |
| I0510 16:00:44.691012 10926 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0612245 | |
| I0510 16:00:44.691033 10926 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.125 | |
| I0510 16:00:44.691046 10926 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.125 | |
| I0510 16:00:44.691059 10926 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.125 | |
| I0510 16:00:44.691071 10926 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0 | |
| I0510 16:00:44.691083 10926 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.375 | |
| I0510 16:00:44.691095 10926 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.375 | |
| I0510 16:00:44.691107 10926 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.75 | |
| I0510 16:00:44.691121 10926 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.75 | |
| I0510 16:00:44.691133 10926 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.875 | |
| I0510 16:00:44.691145 10926 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.875 | |
| I0510 16:00:44.691162 10926 solver.cpp:245] Train net output #11: loss1/accuracy11 = 0.875 | |
| I0510 16:00:44.691174 10926 solver.cpp:245] Train net output #12: loss1/accuracy12 = 0.875 | |
| I0510 16:00:44.691186 10926 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0510 16:00:44.691198 10926 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0510 16:00:44.691213 10926 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0510 16:00:44.691226 10926 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0510 16:00:44.691238 10926 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0510 16:00:44.691251 10926 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0510 16:00:44.691262 10926 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0510 16:00:44.691273 10926 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0510 16:00:44.691285 10926 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0510 16:00:44.691298 10926 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0510 16:00:44.691308 10926 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.732955 | |
| I0510 16:00:44.691324 10926 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.142857 | |
| I0510 16:00:44.691340 10926 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.59161 (* 0.3 = 1.07748 loss) | |
| I0510 16:00:44.691355 10926 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.24167 (* 0.3 = 0.3725 loss) | |
| I0510 16:00:44.691370 10926 solver.cpp:245] Train net output #27: loss1/loss01 = 3.63605 (* 0.0272727 = 0.0991649 loss) | |
| I0510 16:00:44.691385 10926 solver.cpp:245] Train net output #28: loss1/loss02 = 4.03212 (* 0.0272727 = 0.109967 loss) | |
| I0510 16:00:44.691398 10926 solver.cpp:245] Train net output #29: loss1/loss03 = 3.39951 (* 0.0272727 = 0.0927138 loss) | |
| I0510 16:00:44.691412 10926 solver.cpp:245] Train net output #30: loss1/loss04 = 4.03595 (* 0.0272727 = 0.110071 loss) | |
| I0510 16:00:44.691426 10926 solver.cpp:245] Train net output #31: loss1/loss05 = 2.98604 (* 0.0272727 = 0.0814375 loss) | |
| I0510 16:00:44.691440 10926 solver.cpp:245] Train net output #32: loss1/loss06 = 2.32071 (* 0.0272727 = 0.063292 loss) | |
| I0510 16:00:44.691455 10926 solver.cpp:245] Train net output #33: loss1/loss07 = 1.40876 (* 0.0272727 = 0.0384207 loss) | |
| I0510 16:00:44.691469 10926 solver.cpp:245] Train net output #34: loss1/loss08 = 1.21281 (* 0.0272727 = 0.0330767 loss) | |
| I0510 16:00:44.691483 10926 solver.cpp:245] Train net output #35: loss1/loss09 = 0.981781 (* 0.0272727 = 0.0267758 loss) | |
| I0510 16:00:44.691498 10926 solver.cpp:245] Train net output #36: loss1/loss10 = 0.892022 (* 0.0272727 = 0.0243279 loss) | |
| I0510 16:00:44.691512 10926 solver.cpp:245] Train net output #37: loss1/loss11 = 0.716894 (* 0.0272727 = 0.0195517 loss) | |
| I0510 16:00:44.691526 10926 solver.cpp:245] Train net output #38: loss1/loss12 = 0.814973 (* 0.0272727 = 0.0222265 loss) | |
| I0510 16:00:44.691541 10926 solver.cpp:245] Train net output #39: loss1/loss13 = 0.052657 (* 0.0272727 = 0.0014361 loss) | |
| I0510 16:00:44.691582 10926 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0370699 (* 0.0272727 = 0.001011 loss) | |
| I0510 16:00:44.691599 10926 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0273529 (* 0.0272727 = 0.000745989 loss) | |
| I0510 16:00:44.691613 10926 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0288381 (* 0.0272727 = 0.000786495 loss) | |
| I0510 16:00:44.691628 10926 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0108177 (* 0.0272727 = 0.000295027 loss) | |
| I0510 16:00:44.691642 10926 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0105108 (* 0.0272727 = 0.000286657 loss) | |
| I0510 16:00:44.691656 10926 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0141264 (* 0.0272727 = 0.000385266 loss) | |
| I0510 16:00:44.691670 10926 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00789156 (* 0.0272727 = 0.000215224 loss) | |
| I0510 16:00:44.691685 10926 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00760459 (* 0.0272727 = 0.000207398 loss) | |
| I0510 16:00:44.691699 10926 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0166603 (* 0.0272727 = 0.000454373 loss) | |
| I0510 16:00:44.691712 10926 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0816327 | |
| I0510 16:00:44.691725 10926 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.125 | |
| I0510 16:00:44.691736 10926 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0.125 | |
| I0510 16:00:44.691748 10926 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0.125 | |
| I0510 16:00:44.691761 10926 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0 | |
| I0510 16:00:44.691772 10926 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.375 | |
| I0510 16:00:44.691784 10926 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.5 | |
| I0510 16:00:44.691797 10926 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.75 | |
| I0510 16:00:44.691808 10926 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.75 | |
| I0510 16:00:44.691820 10926 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.875 | |
| I0510 16:00:44.691831 10926 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.875 | |
| I0510 16:00:44.691844 10926 solver.cpp:245] Train net output #60: loss2/accuracy11 = 0.875 | |
| I0510 16:00:44.691856 10926 solver.cpp:245] Train net output #61: loss2/accuracy12 = 0.875 | |
| I0510 16:00:44.691867 10926 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0510 16:00:44.691884 10926 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0510 16:00:44.691895 10926 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0510 16:00:44.691907 10926 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0510 16:00:44.691920 10926 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0510 16:00:44.691931 10926 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0510 16:00:44.691942 10926 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0510 16:00:44.691953 10926 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0510 16:00:44.691965 10926 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0510 16:00:44.691977 10926 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0510 16:00:44.691988 10926 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.744318 | |
| I0510 16:00:44.691999 10926 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.183673 | |
| I0510 16:00:44.692013 10926 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.64734 (* 0.3 = 1.0942 loss) | |
| I0510 16:00:44.692030 10926 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.1479 (* 0.3 = 0.34437 loss) | |
| I0510 16:00:44.692045 10926 solver.cpp:245] Train net output #76: loss2/loss01 = 4.04546 (* 0.0272727 = 0.110331 loss) | |
| I0510 16:00:44.692059 10926 solver.cpp:245] Train net output #77: loss2/loss02 = 3.46848 (* 0.0272727 = 0.0945949 loss) | |
| I0510 16:00:44.692085 10926 solver.cpp:245] Train net output #78: loss2/loss03 = 3.45419 (* 0.0272727 = 0.0942052 loss) | |
| I0510 16:00:44.692101 10926 solver.cpp:245] Train net output #79: loss2/loss04 = 3.89581 (* 0.0272727 = 0.106249 loss) | |
| I0510 16:00:44.692116 10926 solver.cpp:245] Train net output #80: loss2/loss05 = 2.91855 (* 0.0272727 = 0.0795968 loss) | |
| I0510 16:00:44.692137 10926 solver.cpp:245] Train net output #81: loss2/loss06 = 2.4409 (* 0.0272727 = 0.0665699 loss) | |
| I0510 16:00:44.692149 10926 solver.cpp:245] Train net output #82: loss2/loss07 = 1.29319 (* 0.0272727 = 0.0352689 loss) | |
| I0510 16:00:44.692163 10926 solver.cpp:245] Train net output #83: loss2/loss08 = 1.30767 (* 0.0272727 = 0.0356638 loss) | |
| I0510 16:00:44.692176 10926 solver.cpp:245] Train net output #84: loss2/loss09 = 0.750678 (* 0.0272727 = 0.020473 loss) | |
| I0510 16:00:44.692191 10926 solver.cpp:245] Train net output #85: loss2/loss10 = 0.640379 (* 0.0272727 = 0.0174649 loss) | |
| I0510 16:00:44.692212 10926 solver.cpp:245] Train net output #86: loss2/loss11 = 0.53993 (* 0.0272727 = 0.0147254 loss) | |
| I0510 16:00:44.692226 10926 solver.cpp:245] Train net output #87: loss2/loss12 = 0.566364 (* 0.0272727 = 0.0154463 loss) | |
| I0510 16:00:44.692240 10926 solver.cpp:245] Train net output #88: loss2/loss13 = 0.0995508 (* 0.0272727 = 0.00271502 loss) | |
| I0510 16:00:44.692255 10926 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0400429 (* 0.0272727 = 0.00109208 loss) | |
| I0510 16:00:44.692268 10926 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0272631 (* 0.0272727 = 0.000743539 loss) | |
| I0510 16:00:44.692282 10926 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0284149 (* 0.0272727 = 0.000774951 loss) | |
| I0510 16:00:44.692296 10926 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0112267 (* 0.0272727 = 0.000306182 loss) | |
| I0510 16:00:44.692311 10926 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0139437 (* 0.0272727 = 0.000380282 loss) | |
| I0510 16:00:44.692324 10926 solver.cpp:245] Train net output #94: loss2/loss19 = 0.00769487 (* 0.0272727 = 0.00020986 loss) | |
| I0510 16:00:44.692339 10926 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00980231 (* 0.0272727 = 0.000267336 loss) | |
| I0510 16:00:44.692353 10926 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00766271 (* 0.0272727 = 0.000208983 loss) | |
| I0510 16:00:44.692368 10926 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0104517 (* 0.0272727 = 0.000285047 loss) | |
| I0510 16:00:44.692379 10926 solver.cpp:245] Train net output #98: loss3/accuracy = 0.142857 | |
| I0510 16:00:44.692391 10926 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.125 | |
| I0510 16:00:44.692404 10926 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.125 | |
| I0510 16:00:44.692415 10926 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0.125 | |
| I0510 16:00:44.692427 10926 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0510 16:00:44.692438 10926 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.375 | |
| I0510 16:00:44.692451 10926 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.625 | |
| I0510 16:00:44.692462 10926 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.75 | |
| I0510 16:00:44.692473 10926 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.75 | |
| I0510 16:00:44.692486 10926 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.875 | |
| I0510 16:00:44.692497 10926 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.875 | |
| I0510 16:00:44.692508 10926 solver.cpp:245] Train net output #109: loss3/accuracy11 = 0.875 | |
| I0510 16:00:44.692520 10926 solver.cpp:245] Train net output #110: loss3/accuracy12 = 0.875 | |
| I0510 16:00:44.692533 10926 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0510 16:00:44.692543 10926 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0510 16:00:44.692555 10926 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0510 16:00:44.692566 10926 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0510 16:00:44.692589 10926 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0510 16:00:44.692605 10926 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0510 16:00:44.692618 10926 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0510 16:00:44.692631 10926 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0510 16:00:44.692641 10926 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0510 16:00:44.692653 10926 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0510 16:00:44.692664 10926 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.75 | |
| I0510 16:00:44.692677 10926 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.204082 | |
| I0510 16:00:44.692697 10926 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.64686 (* 1 = 3.64686 loss) | |
| I0510 16:00:44.692710 10926 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.21841 (* 1 = 1.21841 loss) | |
| I0510 16:00:44.692723 10926 solver.cpp:245] Train net output #125: loss3/loss01 = 3.79627 (* 0.0909091 = 0.345116 loss) | |
| I0510 16:00:44.692737 10926 solver.cpp:245] Train net output #126: loss3/loss02 = 3.61644 (* 0.0909091 = 0.328767 loss) | |
| I0510 16:00:44.692750 10926 solver.cpp:245] Train net output #127: loss3/loss03 = 3.49699 (* 0.0909091 = 0.317908 loss) | |
| I0510 16:00:44.692773 10926 solver.cpp:245] Train net output #128: loss3/loss04 = 3.80086 (* 0.0909091 = 0.345533 loss) | |
| I0510 16:00:44.692787 10926 solver.cpp:245] Train net output #129: loss3/loss05 = 2.94571 (* 0.0909091 = 0.267792 loss) | |
| I0510 16:00:44.692800 10926 solver.cpp:245] Train net output #130: loss3/loss06 = 2.07153 (* 0.0909091 = 0.188321 loss) | |
| I0510 16:00:44.692814 10926 solver.cpp:245] Train net output #131: loss3/loss07 = 1.2075 (* 0.0909091 = 0.109773 loss) | |
| I0510 16:00:44.692828 10926 solver.cpp:245] Train net output #132: loss3/loss08 = 1.12545 (* 0.0909091 = 0.102314 loss) | |
| I0510 16:00:44.692842 10926 solver.cpp:245] Train net output #133: loss3/loss09 = 0.660811 (* 0.0909091 = 0.0600737 loss) | |
| I0510 16:00:44.692855 10926 solver.cpp:245] Train net output #134: loss3/loss10 = 0.448927 (* 0.0909091 = 0.0408116 loss) | |
| I0510 16:00:44.692869 10926 solver.cpp:245] Train net output #135: loss3/loss11 = 0.43072 (* 0.0909091 = 0.0391564 loss) | |
| I0510 16:00:44.692883 10926 solver.cpp:245] Train net output #136: loss3/loss12 = 0.328807 (* 0.0909091 = 0.0298916 loss) | |
| I0510 16:00:44.692898 10926 solver.cpp:245] Train net output #137: loss3/loss13 = 0.0655358 (* 0.0909091 = 0.0059578 loss) | |
| I0510 16:00:44.692911 10926 solver.cpp:245] Train net output #138: loss3/loss14 = 0.0694435 (* 0.0909091 = 0.00631305 loss) | |
| I0510 16:00:44.692927 10926 solver.cpp:245] Train net output #139: loss3/loss15 = 0.0351247 (* 0.0909091 = 0.00319315 loss) | |
| I0510 16:00:44.692942 10926 solver.cpp:245] Train net output #140: loss3/loss16 = 0.0193202 (* 0.0909091 = 0.00175638 loss) | |
| I0510 16:00:44.692957 10926 solver.cpp:245] Train net output #141: loss3/loss17 = 0.0119852 (* 0.0909091 = 0.00108956 loss) | |
| I0510 16:00:44.692970 10926 solver.cpp:245] Train net output #142: loss3/loss18 = 0.0070514 (* 0.0909091 = 0.000641036 loss) | |
| I0510 16:00:44.692984 10926 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00226859 (* 0.0909091 = 0.000206235 loss) | |
| I0510 16:00:44.692998 10926 solver.cpp:245] Train net output #144: loss3/loss20 = 0.0029497 (* 0.0909091 = 0.000268155 loss) | |
| I0510 16:00:44.693012 10926 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00167 (* 0.0909091 = 0.000151819 loss) | |
| I0510 16:00:44.693027 10926 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00154466 (* 0.0909091 = 0.000140424 loss) | |
| I0510 16:00:44.693038 10926 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0510 16:00:44.693049 10926 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0510 16:00:44.693061 10926 solver.cpp:245] Train net output #149: total_confidence = 3.50662e-06 | |
| I0510 16:00:44.693097 10926 solver.cpp:245] Train net output #150: total_confidence_not_rec = 5.35683e-05 | |
| I0510 16:00:44.693114 10926 sgd_solver.cpp:106] Iteration 7000, lr = 0.001 | |
| I0510 16:01:23.206826 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 35.0996 > 30) by scale factor 0.854711 | |
| I0510 16:01:49.457360 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 65.4402 > 30) by scale factor 0.458433 | |
| I0510 16:02:03.900523 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 30.2761 > 30) by scale factor 0.990882 | |
| I0510 16:02:13.356412 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 32.8503 > 30) by scale factor 0.913233 | |
| I0510 16:02:16.603520 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 45.2418 > 30) by scale factor 0.663104 | |
| I0510 16:02:37.282879 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 33.1938 > 30) by scale factor 0.903784 | |
| I0510 16:02:50.880519 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 35.2284 > 30) by scale factor 0.851585 | |
| I0510 16:03:12.342713 10926 solver.cpp:229] Iteration 7500, loss = 11.0623 | |
| I0510 16:03:12.342871 10926 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0697674 | |
| I0510 16:03:12.342906 10926 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.125 | |
| I0510 16:03:12.342942 10926 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.125 | |
| I0510 16:03:12.342969 10926 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0510 16:03:12.342993 10926 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.25 | |
| I0510 16:03:12.343015 10926 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.25 | |
| I0510 16:03:12.343029 10926 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.5 | |
| I0510 16:03:12.343040 10926 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.625 | |
| I0510 16:03:12.343053 10926 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.75 | |
| I0510 16:03:12.343071 10926 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0510 16:03:12.343083 10926 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0510 16:03:12.343096 10926 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0510 16:03:12.343108 10926 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0510 16:03:12.343122 10926 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0510 16:03:12.343138 10926 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0510 16:03:12.343149 10926 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0510 16:03:12.343161 10926 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0510 16:03:12.343173 10926 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0510 16:03:12.343185 10926 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0510 16:03:12.343199 10926 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0510 16:03:12.343211 10926 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0510 16:03:12.343224 10926 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0510 16:03:12.343235 10926 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0510 16:03:12.343246 10926 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.761364 | |
| I0510 16:03:12.343258 10926 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.209302 | |
| I0510 16:03:12.343274 10926 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.52528 (* 0.3 = 1.05758 loss) | |
| I0510 16:03:12.343289 10926 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.10652 (* 0.3 = 0.331955 loss) | |
| I0510 16:03:12.343304 10926 solver.cpp:245] Train net output #27: loss1/loss01 = 3.75639 (* 0.0272727 = 0.102447 loss) | |
| I0510 16:03:12.343318 10926 solver.cpp:245] Train net output #28: loss1/loss02 = 3.33747 (* 0.0272727 = 0.0910218 loss) | |
| I0510 16:03:12.343333 10926 solver.cpp:245] Train net output #29: loss1/loss03 = 4.45098 (* 0.0272727 = 0.12139 loss) | |
| I0510 16:03:12.343348 10926 solver.cpp:245] Train net output #30: loss1/loss04 = 3.33364 (* 0.0272727 = 0.0909174 loss) | |
| I0510 16:03:12.343361 10926 solver.cpp:245] Train net output #31: loss1/loss05 = 3.24684 (* 0.0272727 = 0.0885502 loss) | |
| I0510 16:03:12.343375 10926 solver.cpp:245] Train net output #32: loss1/loss06 = 2.45267 (* 0.0272727 = 0.066891 loss) | |
| I0510 16:03:12.343389 10926 solver.cpp:245] Train net output #33: loss1/loss07 = 2.12699 (* 0.0272727 = 0.0580088 loss) | |
| I0510 16:03:12.343403 10926 solver.cpp:245] Train net output #34: loss1/loss08 = 1.95299 (* 0.0272727 = 0.0532633 loss) | |
| I0510 16:03:12.343417 10926 solver.cpp:245] Train net output #35: loss1/loss09 = 0.183177 (* 0.0272727 = 0.00499574 loss) | |
| I0510 16:03:12.343431 10926 solver.cpp:245] Train net output #36: loss1/loss10 = 0.157201 (* 0.0272727 = 0.00428731 loss) | |
| I0510 16:03:12.343446 10926 solver.cpp:245] Train net output #37: loss1/loss11 = 0.102302 (* 0.0272727 = 0.00279006 loss) | |
| I0510 16:03:12.343461 10926 solver.cpp:245] Train net output #38: loss1/loss12 = 0.0940255 (* 0.0272727 = 0.00256433 loss) | |
| I0510 16:03:12.343475 10926 solver.cpp:245] Train net output #39: loss1/loss13 = 0.0798285 (* 0.0272727 = 0.00217714 loss) | |
| I0510 16:03:12.343716 10926 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0575055 (* 0.0272727 = 0.00156833 loss) | |
| I0510 16:03:12.343745 10926 solver.cpp:245] Train net output #41: loss1/loss15 = 0.03768 (* 0.0272727 = 0.00102764 loss) | |
| I0510 16:03:12.343761 10926 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0255905 (* 0.0272727 = 0.000697924 loss) | |
| I0510 16:03:12.343776 10926 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0372715 (* 0.0272727 = 0.00101649 loss) | |
| I0510 16:03:12.343791 10926 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0226236 (* 0.0272727 = 0.000617007 loss) | |
| I0510 16:03:12.343806 10926 solver.cpp:245] Train net output #45: loss1/loss19 = 0.020982 (* 0.0272727 = 0.000572236 loss) | |
| I0510 16:03:12.343829 10926 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0156744 (* 0.0272727 = 0.000427483 loss) | |
| I0510 16:03:12.343843 10926 solver.cpp:245] Train net output #47: loss1/loss21 = 0.017723 (* 0.0272727 = 0.000483354 loss) | |
| I0510 16:03:12.343858 10926 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0265294 (* 0.0272727 = 0.000723528 loss) | |
| I0510 16:03:12.343871 10926 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0697674 | |
| I0510 16:03:12.343889 10926 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0510 16:03:12.343902 10926 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0510 16:03:12.343914 10926 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0.25 | |
| I0510 16:03:12.343926 10926 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.25 | |
| I0510 16:03:12.343938 10926 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.375 | |
| I0510 16:03:12.343950 10926 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.625 | |
| I0510 16:03:12.343962 10926 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.625 | |
| I0510 16:03:12.343974 10926 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.75 | |
| I0510 16:03:12.343986 10926 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0510 16:03:12.343998 10926 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0510 16:03:12.344010 10926 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0510 16:03:12.344022 10926 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0510 16:03:12.344033 10926 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0510 16:03:12.344044 10926 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0510 16:03:12.344056 10926 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0510 16:03:12.344069 10926 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0510 16:03:12.344080 10926 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0510 16:03:12.344091 10926 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0510 16:03:12.344104 10926 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0510 16:03:12.344115 10926 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0510 16:03:12.344126 10926 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0510 16:03:12.344137 10926 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0510 16:03:12.344148 10926 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.744318 | |
| I0510 16:03:12.344161 10926 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.116279 | |
| I0510 16:03:12.344175 10926 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.43308 (* 0.3 = 1.02992 loss) | |
| I0510 16:03:12.344190 10926 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.15921 (* 0.3 = 0.347762 loss) | |
| I0510 16:03:12.344204 10926 solver.cpp:245] Train net output #76: loss2/loss01 = 3.66346 (* 0.0272727 = 0.0999126 loss) | |
| I0510 16:03:12.344218 10926 solver.cpp:245] Train net output #77: loss2/loss02 = 3.62073 (* 0.0272727 = 0.0987472 loss) | |
| I0510 16:03:12.344245 10926 solver.cpp:245] Train net output #78: loss2/loss03 = 4.02778 (* 0.0272727 = 0.109849 loss) | |
| I0510 16:03:12.344264 10926 solver.cpp:245] Train net output #79: loss2/loss04 = 3.02545 (* 0.0272727 = 0.0825124 loss) | |
| I0510 16:03:12.344291 10926 solver.cpp:245] Train net output #80: loss2/loss05 = 2.95014 (* 0.0272727 = 0.0804583 loss) | |
| I0510 16:03:12.344317 10926 solver.cpp:245] Train net output #81: loss2/loss06 = 2.40827 (* 0.0272727 = 0.0656802 loss) | |
| I0510 16:03:12.344332 10926 solver.cpp:245] Train net output #82: loss2/loss07 = 2.1246 (* 0.0272727 = 0.0579436 loss) | |
| I0510 16:03:12.344347 10926 solver.cpp:245] Train net output #83: loss2/loss08 = 2.14434 (* 0.0272727 = 0.0584819 loss) | |
| I0510 16:03:12.344362 10926 solver.cpp:245] Train net output #84: loss2/loss09 = 0.197573 (* 0.0272727 = 0.00538834 loss) | |
| I0510 16:03:12.344375 10926 solver.cpp:245] Train net output #85: loss2/loss10 = 0.106007 (* 0.0272727 = 0.00289111 loss) | |
| I0510 16:03:12.344389 10926 solver.cpp:245] Train net output #86: loss2/loss11 = 0.0706577 (* 0.0272727 = 0.00192703 loss) | |
| I0510 16:03:12.344403 10926 solver.cpp:245] Train net output #87: loss2/loss12 = 0.0776131 (* 0.0272727 = 0.00211672 loss) | |
| I0510 16:03:12.344419 10926 solver.cpp:245] Train net output #88: loss2/loss13 = 0.0790247 (* 0.0272727 = 0.00215522 loss) | |
| I0510 16:03:12.344429 10926 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0450661 (* 0.0272727 = 0.00122908 loss) | |
| I0510 16:03:12.344439 10926 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0572529 (* 0.0272727 = 0.00156144 loss) | |
| I0510 16:03:12.344454 10926 solver.cpp:245] Train net output #91: loss2/loss16 = 0.014083 (* 0.0272727 = 0.000384081 loss) | |
| I0510 16:03:12.344467 10926 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0177451 (* 0.0272727 = 0.000483957 loss) | |
| I0510 16:03:12.344481 10926 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0185111 (* 0.0272727 = 0.000504849 loss) | |
| I0510 16:03:12.344496 10926 solver.cpp:245] Train net output #94: loss2/loss19 = 0.0482827 (* 0.0272727 = 0.0013168 loss) | |
| I0510 16:03:12.344511 10926 solver.cpp:245] Train net output #95: loss2/loss20 = 0.0176419 (* 0.0272727 = 0.000481142 loss) | |
| I0510 16:03:12.344524 10926 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00815013 (* 0.0272727 = 0.000222276 loss) | |
| I0510 16:03:12.344538 10926 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0218135 (* 0.0272727 = 0.000594914 loss) | |
| I0510 16:03:12.344552 10926 solver.cpp:245] Train net output #98: loss3/accuracy = 0.116279 | |
| I0510 16:03:12.344563 10926 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.125 | |
| I0510 16:03:12.344576 10926 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.125 | |
| I0510 16:03:12.344588 10926 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0510 16:03:12.344599 10926 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.25 | |
| I0510 16:03:12.344611 10926 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.25 | |
| I0510 16:03:12.344624 10926 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.5 | |
| I0510 16:03:12.344635 10926 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.5 | |
| I0510 16:03:12.344647 10926 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.75 | |
| I0510 16:03:12.344658 10926 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0510 16:03:12.344671 10926 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0510 16:03:12.344681 10926 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0510 16:03:12.344704 10926 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0510 16:03:12.344717 10926 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0510 16:03:12.344729 10926 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0510 16:03:12.344740 10926 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0510 16:03:12.344753 10926 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0510 16:03:12.344774 10926 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0510 16:03:12.344787 10926 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0510 16:03:12.344799 10926 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0510 16:03:12.344810 10926 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0510 16:03:12.344821 10926 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0510 16:03:12.344833 10926 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0510 16:03:12.344844 10926 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.772727 | |
| I0510 16:03:12.344856 10926 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.348837 | |
| I0510 16:03:12.344871 10926 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.07827 (* 1 = 3.07827 loss) | |
| I0510 16:03:12.344884 10926 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 0.90439 (* 1 = 0.90439 loss) | |
| I0510 16:03:12.344897 10926 solver.cpp:245] Train net output #125: loss3/loss01 = 3.07139 (* 0.0909091 = 0.279217 loss) | |
| I0510 16:03:12.344913 10926 solver.cpp:245] Train net output #126: loss3/loss02 = 2.85249 (* 0.0909091 = 0.259317 loss) | |
| I0510 16:03:12.344928 10926 solver.cpp:245] Train net output #127: loss3/loss03 = 3.80708 (* 0.0909091 = 0.346098 loss) | |
| I0510 16:03:12.344944 10926 solver.cpp:245] Train net output #128: loss3/loss04 = 2.47268 (* 0.0909091 = 0.224789 loss) | |
| I0510 16:03:12.344957 10926 solver.cpp:245] Train net output #129: loss3/loss05 = 2.86616 (* 0.0909091 = 0.26056 loss) | |
| I0510 16:03:12.344971 10926 solver.cpp:245] Train net output #130: loss3/loss06 = 1.83111 (* 0.0909091 = 0.166465 loss) | |
| I0510 16:03:12.344985 10926 solver.cpp:245] Train net output #131: loss3/loss07 = 1.68562 (* 0.0909091 = 0.153238 loss) | |
| I0510 16:03:12.345000 10926 solver.cpp:245] Train net output #132: loss3/loss08 = 1.76957 (* 0.0909091 = 0.16087 loss) | |
| I0510 16:03:12.345013 10926 solver.cpp:245] Train net output #133: loss3/loss09 = 0.0898332 (* 0.0909091 = 0.00816665 loss) | |
| I0510 16:03:12.345027 10926 solver.cpp:245] Train net output #134: loss3/loss10 = 0.0411102 (* 0.0909091 = 0.00373729 loss) | |
| I0510 16:03:12.345041 10926 solver.cpp:245] Train net output #135: loss3/loss11 = 0.0265159 (* 0.0909091 = 0.00241053 loss) | |
| I0510 16:03:12.345055 10926 solver.cpp:245] Train net output #136: loss3/loss12 = 0.0304685 (* 0.0909091 = 0.00276986 loss) | |
| I0510 16:03:12.345070 10926 solver.cpp:245] Train net output #137: loss3/loss13 = 0.020528 (* 0.0909091 = 0.00186618 loss) | |
| I0510 16:03:12.345084 10926 solver.cpp:245] Train net output #138: loss3/loss14 = 0.0221844 (* 0.0909091 = 0.00201676 loss) | |
| I0510 16:03:12.345098 10926 solver.cpp:245] Train net output #139: loss3/loss15 = 0.0282553 (* 0.0909091 = 0.00256867 loss) | |
| I0510 16:03:12.345134 10926 solver.cpp:245] Train net output #140: loss3/loss16 = 0.0191102 (* 0.0909091 = 0.00173729 loss) | |
| I0510 16:03:12.345151 10926 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00780239 (* 0.0909091 = 0.000709308 loss) | |
| I0510 16:03:12.345173 10926 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00811076 (* 0.0909091 = 0.000737342 loss) | |
| I0510 16:03:12.345187 10926 solver.cpp:245] Train net output #143: loss3/loss19 = 0.0078958 (* 0.0909091 = 0.0007178 loss) | |
| I0510 16:03:12.345202 10926 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00261993 (* 0.0909091 = 0.000238175 loss) | |
| I0510 16:03:12.345216 10926 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00153594 (* 0.0909091 = 0.000139631 loss) | |
| I0510 16:03:12.345229 10926 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00121079 (* 0.0909091 = 0.000110072 loss) | |
| I0510 16:03:12.345242 10926 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0510 16:03:12.345254 10926 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0510 16:03:12.345266 10926 solver.cpp:245] Train net output #149: total_confidence = 3.76076e-05 | |
| I0510 16:03:12.345290 10926 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.000656891 | |
| I0510 16:03:12.345305 10926 sgd_solver.cpp:106] Iteration 7500, lr = 0.001 | |
| I0510 16:05:04.240294 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 37.7585 > 30) by scale factor 0.794523 | |
| I0510 16:05:10.715296 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 69.3949 > 30) by scale factor 0.432309 | |
| I0510 16:05:18.655457 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 30.8968 > 30) by scale factor 0.970975 | |
| I0510 16:05:33.081733 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 31.5803 > 30) by scale factor 0.94996 | |
| I0510 16:05:39.427021 10926 solver.cpp:229] Iteration 8000, loss = 10.9712 | |
| I0510 16:05:39.427165 10926 solver.cpp:245] Train net output #0: loss1/accuracy = 0.046875 | |
| I0510 16:05:39.427187 10926 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.125 | |
| I0510 16:05:39.427201 10926 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0510 16:05:39.427214 10926 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0510 16:05:39.427227 10926 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.125 | |
| I0510 16:05:39.427238 10926 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.125 | |
| I0510 16:05:39.427250 10926 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.125 | |
| I0510 16:05:39.427263 10926 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.625 | |
| I0510 16:05:39.427276 10926 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.75 | |
| I0510 16:05:39.427289 10926 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.875 | |
| I0510 16:05:39.427300 10926 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.75 | |
| I0510 16:05:39.427314 10926 solver.cpp:245] Train net output #11: loss1/accuracy11 = 0.75 | |
| I0510 16:05:39.427325 10926 solver.cpp:245] Train net output #12: loss1/accuracy12 = 0.875 | |
| I0510 16:05:39.427337 10926 solver.cpp:245] Train net output #13: loss1/accuracy13 = 0.875 | |
| I0510 16:05:39.427350 10926 solver.cpp:245] Train net output #14: loss1/accuracy14 = 0.875 | |
| I0510 16:05:39.427362 10926 solver.cpp:245] Train net output #15: loss1/accuracy15 = 0.875 | |
| I0510 16:05:39.427374 10926 solver.cpp:245] Train net output #16: loss1/accuracy16 = 0.875 | |
| I0510 16:05:39.427386 10926 solver.cpp:245] Train net output #17: loss1/accuracy17 = 0.875 | |
| I0510 16:05:39.427399 10926 solver.cpp:245] Train net output #18: loss1/accuracy18 = 0.875 | |
| I0510 16:05:39.427412 10926 solver.cpp:245] Train net output #19: loss1/accuracy19 = 0.875 | |
| I0510 16:05:39.427423 10926 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0510 16:05:39.427435 10926 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0510 16:05:39.427448 10926 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0510 16:05:39.427459 10926 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.647727 | |
| I0510 16:05:39.427471 10926 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.125 | |
| I0510 16:05:39.427489 10926 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.5431 (* 0.3 = 1.06293 loss) | |
| I0510 16:05:39.427503 10926 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.51382 (* 0.3 = 0.454147 loss) | |
| I0510 16:05:39.427518 10926 solver.cpp:245] Train net output #27: loss1/loss01 = 3.46792 (* 0.0272727 = 0.0945796 loss) | |
| I0510 16:05:39.427532 10926 solver.cpp:245] Train net output #28: loss1/loss02 = 3.85346 (* 0.0272727 = 0.105094 loss) | |
| I0510 16:05:39.427546 10926 solver.cpp:245] Train net output #29: loss1/loss03 = 3.76325 (* 0.0272727 = 0.102634 loss) | |
| I0510 16:05:39.427561 10926 solver.cpp:245] Train net output #30: loss1/loss04 = 3.6074 (* 0.0272727 = 0.0983836 loss) | |
| I0510 16:05:39.427574 10926 solver.cpp:245] Train net output #31: loss1/loss05 = 3.65175 (* 0.0272727 = 0.0995931 loss) | |
| I0510 16:05:39.427589 10926 solver.cpp:245] Train net output #32: loss1/loss06 = 3.57996 (* 0.0272727 = 0.0976351 loss) | |
| I0510 16:05:39.427603 10926 solver.cpp:245] Train net output #33: loss1/loss07 = 2.25765 (* 0.0272727 = 0.0615722 loss) | |
| I0510 16:05:39.427618 10926 solver.cpp:245] Train net output #34: loss1/loss08 = 1.10329 (* 0.0272727 = 0.0300897 loss) | |
| I0510 16:05:39.427633 10926 solver.cpp:245] Train net output #35: loss1/loss09 = 0.642292 (* 0.0272727 = 0.0175171 loss) | |
| I0510 16:05:39.427647 10926 solver.cpp:245] Train net output #36: loss1/loss10 = 1.42307 (* 0.0272727 = 0.0388109 loss) | |
| I0510 16:05:39.427661 10926 solver.cpp:245] Train net output #37: loss1/loss11 = 1.54242 (* 0.0272727 = 0.0420661 loss) | |
| I0510 16:05:39.427675 10926 solver.cpp:245] Train net output #38: loss1/loss12 = 0.647333 (* 0.0272727 = 0.0176545 loss) | |
| I0510 16:05:39.427711 10926 solver.cpp:245] Train net output #39: loss1/loss13 = 0.761927 (* 0.0272727 = 0.0207798 loss) | |
| I0510 16:05:39.427727 10926 solver.cpp:245] Train net output #40: loss1/loss14 = 0.749522 (* 0.0272727 = 0.0204415 loss) | |
| I0510 16:05:39.427742 10926 solver.cpp:245] Train net output #41: loss1/loss15 = 0.7982 (* 0.0272727 = 0.0217691 loss) | |
| I0510 16:05:39.427757 10926 solver.cpp:245] Train net output #42: loss1/loss16 = 0.810678 (* 0.0272727 = 0.0221094 loss) | |
| I0510 16:05:39.427772 10926 solver.cpp:245] Train net output #43: loss1/loss17 = 0.930548 (* 0.0272727 = 0.0253786 loss) | |
| I0510 16:05:39.427785 10926 solver.cpp:245] Train net output #44: loss1/loss18 = 0.975409 (* 0.0272727 = 0.0266021 loss) | |
| I0510 16:05:39.427799 10926 solver.cpp:245] Train net output #45: loss1/loss19 = 0.922455 (* 0.0272727 = 0.0251579 loss) | |
| I0510 16:05:39.427815 10926 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0293871 (* 0.0272727 = 0.000801467 loss) | |
| I0510 16:05:39.427829 10926 solver.cpp:245] Train net output #47: loss1/loss21 = 0.015549 (* 0.0272727 = 0.000424064 loss) | |
| I0510 16:05:39.427845 10926 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0202247 (* 0.0272727 = 0.000551583 loss) | |
| I0510 16:05:39.427857 10926 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0625 | |
| I0510 16:05:39.427870 10926 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0510 16:05:39.427884 10926 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0510 16:05:39.427897 10926 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0510 16:05:39.427909 10926 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.25 | |
| I0510 16:05:39.427922 10926 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.25 | |
| I0510 16:05:39.427934 10926 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.125 | |
| I0510 16:05:39.427947 10926 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.625 | |
| I0510 16:05:39.427958 10926 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.75 | |
| I0510 16:05:39.427970 10926 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.75 | |
| I0510 16:05:39.427983 10926 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.75 | |
| I0510 16:05:39.427995 10926 solver.cpp:245] Train net output #60: loss2/accuracy11 = 0.75 | |
| I0510 16:05:39.428006 10926 solver.cpp:245] Train net output #61: loss2/accuracy12 = 0.875 | |
| I0510 16:05:39.428020 10926 solver.cpp:245] Train net output #62: loss2/accuracy13 = 0.875 | |
| I0510 16:05:39.428031 10926 solver.cpp:245] Train net output #63: loss2/accuracy14 = 0.875 | |
| I0510 16:05:39.428043 10926 solver.cpp:245] Train net output #64: loss2/accuracy15 = 0.875 | |
| I0510 16:05:39.428056 10926 solver.cpp:245] Train net output #65: loss2/accuracy16 = 0.875 | |
| I0510 16:05:39.428067 10926 solver.cpp:245] Train net output #66: loss2/accuracy17 = 0.875 | |
| I0510 16:05:39.428079 10926 solver.cpp:245] Train net output #67: loss2/accuracy18 = 0.875 | |
| I0510 16:05:39.428092 10926 solver.cpp:245] Train net output #68: loss2/accuracy19 = 0.875 | |
| I0510 16:05:39.428104 10926 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0510 16:05:39.428115 10926 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0510 16:05:39.428128 10926 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0510 16:05:39.428138 10926 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.659091 | |
| I0510 16:05:39.428150 10926 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.171875 | |
| I0510 16:05:39.428165 10926 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.55217 (* 0.3 = 1.06565 loss) | |
| I0510 16:05:39.428184 10926 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.53296 (* 0.3 = 0.459887 loss) | |
| I0510 16:05:39.428200 10926 solver.cpp:245] Train net output #76: loss2/loss01 = 3.69812 (* 0.0272727 = 0.100858 loss) | |
| I0510 16:05:39.428213 10926 solver.cpp:245] Train net output #77: loss2/loss02 = 3.61833 (* 0.0272727 = 0.0986817 loss) | |
| I0510 16:05:39.428241 10926 solver.cpp:245] Train net output #78: loss2/loss03 = 4.25043 (* 0.0272727 = 0.115921 loss) | |
| I0510 16:05:39.428256 10926 solver.cpp:245] Train net output #79: loss2/loss04 = 3.60044 (* 0.0272727 = 0.0981939 loss) | |
| I0510 16:05:39.428269 10926 solver.cpp:245] Train net output #80: loss2/loss05 = 3.31857 (* 0.0272727 = 0.0905065 loss) | |
| I0510 16:05:39.428283 10926 solver.cpp:245] Train net output #81: loss2/loss06 = 3.64432 (* 0.0272727 = 0.0993906 loss) | |
| I0510 16:05:39.428297 10926 solver.cpp:245] Train net output #82: loss2/loss07 = 1.98148 (* 0.0272727 = 0.0540403 loss) | |
| I0510 16:05:39.428311 10926 solver.cpp:245] Train net output #83: loss2/loss08 = 1.01497 (* 0.0272727 = 0.0276809 loss) | |
| I0510 16:05:39.428325 10926 solver.cpp:245] Train net output #84: loss2/loss09 = 0.824002 (* 0.0272727 = 0.0224728 loss) | |
| I0510 16:05:39.428339 10926 solver.cpp:245] Train net output #85: loss2/loss10 = 1.17064 (* 0.0272727 = 0.0319264 loss) | |
| I0510 16:05:39.428354 10926 solver.cpp:245] Train net output #86: loss2/loss11 = 1.2572 (* 0.0272727 = 0.0342872 loss) | |
| I0510 16:05:39.428369 10926 solver.cpp:245] Train net output #87: loss2/loss12 = 0.644292 (* 0.0272727 = 0.0175716 loss) | |
| I0510 16:05:39.428381 10926 solver.cpp:245] Train net output #88: loss2/loss13 = 0.618301 (* 0.0272727 = 0.0168628 loss) | |
| I0510 16:05:39.428396 10926 solver.cpp:245] Train net output #89: loss2/loss14 = 0.711386 (* 0.0272727 = 0.0194014 loss) | |
| I0510 16:05:39.428409 10926 solver.cpp:245] Train net output #90: loss2/loss15 = 0.76764 (* 0.0272727 = 0.0209356 loss) | |
| I0510 16:05:39.428424 10926 solver.cpp:245] Train net output #91: loss2/loss16 = 0.620009 (* 0.0272727 = 0.0169093 loss) | |
| I0510 16:05:39.428438 10926 solver.cpp:245] Train net output #92: loss2/loss17 = 0.661667 (* 0.0272727 = 0.0180455 loss) | |
| I0510 16:05:39.428452 10926 solver.cpp:245] Train net output #93: loss2/loss18 = 0.833619 (* 0.0272727 = 0.0227351 loss) | |
| I0510 16:05:39.428467 10926 solver.cpp:245] Train net output #94: loss2/loss19 = 0.902344 (* 0.0272727 = 0.0246094 loss) | |
| I0510 16:05:39.428480 10926 solver.cpp:245] Train net output #95: loss2/loss20 = 0.0219857 (* 0.0272727 = 0.000599609 loss) | |
| I0510 16:05:39.428494 10926 solver.cpp:245] Train net output #96: loss2/loss21 = 0.0112682 (* 0.0272727 = 0.000307316 loss) | |
| I0510 16:05:39.428508 10926 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0127579 (* 0.0272727 = 0.000347943 loss) | |
| I0510 16:05:39.428519 10926 solver.cpp:245] Train net output #98: loss3/accuracy = 0.09375 | |
| I0510 16:05:39.428527 10926 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0510 16:05:39.428539 10926 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0510 16:05:39.428552 10926 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0.25 | |
| I0510 16:05:39.428565 10926 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0510 16:05:39.428576 10926 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.125 | |
| I0510 16:05:39.428588 10926 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.25 | |
| I0510 16:05:39.428601 10926 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.5 | |
| I0510 16:05:39.428612 10926 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.75 | |
| I0510 16:05:39.428624 10926 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.875 | |
| I0510 16:05:39.428635 10926 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.75 | |
| I0510 16:05:39.428647 10926 solver.cpp:245] Train net output #109: loss3/accuracy11 = 0.75 | |
| I0510 16:05:39.428659 10926 solver.cpp:245] Train net output #110: loss3/accuracy12 = 0.875 | |
| I0510 16:05:39.428670 10926 solver.cpp:245] Train net output #111: loss3/accuracy13 = 0.875 | |
| I0510 16:05:39.428683 10926 solver.cpp:245] Train net output #112: loss3/accuracy14 = 0.875 | |
| I0510 16:05:39.428694 10926 solver.cpp:245] Train net output #113: loss3/accuracy15 = 0.875 | |
| I0510 16:05:39.428705 10926 solver.cpp:245] Train net output #114: loss3/accuracy16 = 0.875 | |
| I0510 16:05:39.428727 10926 solver.cpp:245] Train net output #115: loss3/accuracy17 = 0.875 | |
| I0510 16:05:39.428740 10926 solver.cpp:245] Train net output #116: loss3/accuracy18 = 0.875 | |
| I0510 16:05:39.428753 10926 solver.cpp:245] Train net output #117: loss3/accuracy19 = 0.875 | |
| I0510 16:05:39.428764 10926 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0510 16:05:39.428776 10926 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0510 16:05:39.428788 10926 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0510 16:05:39.428799 10926 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.636364 | |
| I0510 16:05:39.428812 10926 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.234375 | |
| I0510 16:05:39.428825 10926 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.25006 (* 1 = 3.25006 loss) | |
| I0510 16:05:39.428839 10926 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.59899 (* 1 = 1.59899 loss) | |
| I0510 16:05:39.428853 10926 solver.cpp:245] Train net output #125: loss3/loss01 = 3.07483 (* 0.0909091 = 0.27953 loss) | |
| I0510 16:05:39.428867 10926 solver.cpp:245] Train net output #126: loss3/loss02 = 3.28437 (* 0.0909091 = 0.298579 loss) | |
| I0510 16:05:39.428881 10926 solver.cpp:245] Train net output #127: loss3/loss03 = 3.39989 (* 0.0909091 = 0.30908 loss) | |
| I0510 16:05:39.428895 10926 solver.cpp:245] Train net output #128: loss3/loss04 = 3.0681 (* 0.0909091 = 0.278918 loss) | |
| I0510 16:05:39.428910 10926 solver.cpp:245] Train net output #129: loss3/loss05 = 3.32604 (* 0.0909091 = 0.302367 loss) | |
| I0510 16:05:39.428923 10926 solver.cpp:245] Train net output #130: loss3/loss06 = 2.97971 (* 0.0909091 = 0.270883 loss) | |
| I0510 16:05:39.428939 10926 solver.cpp:245] Train net output #131: loss3/loss07 = 2.19637 (* 0.0909091 = 0.19967 loss) | |
| I0510 16:05:39.428953 10926 solver.cpp:245] Train net output #132: loss3/loss08 = 1.14434 (* 0.0909091 = 0.104031 loss) | |
| I0510 16:05:39.428968 10926 solver.cpp:245] Train net output #133: loss3/loss09 = 0.757157 (* 0.0909091 = 0.0688325 loss) | |
| I0510 16:05:39.428982 10926 solver.cpp:245] Train net output #134: loss3/loss10 = 1.07767 (* 0.0909091 = 0.0979699 loss) | |
| I0510 16:05:39.429002 10926 solver.cpp:245] Train net output #135: loss3/loss11 = 0.971137 (* 0.0909091 = 0.0882852 loss) | |
| I0510 16:05:39.429029 10926 solver.cpp:245] Train net output #136: loss3/loss12 = 0.627686 (* 0.0909091 = 0.0570624 loss) | |
| I0510 16:05:39.429049 10926 solver.cpp:245] Train net output #137: loss3/loss13 = 0.565247 (* 0.0909091 = 0.0513861 loss) | |
| I0510 16:05:39.429064 10926 solver.cpp:245] Train net output #138: loss3/loss14 = 0.633778 (* 0.0909091 = 0.0576162 loss) | |
| I0510 16:05:39.429078 10926 solver.cpp:245] Train net output #139: loss3/loss15 = 0.456043 (* 0.0909091 = 0.0414585 loss) | |
| I0510 16:05:39.429092 10926 solver.cpp:245] Train net output #140: loss3/loss16 = 0.783962 (* 0.0909091 = 0.0712693 loss) | |
| I0510 16:05:39.429106 10926 solver.cpp:245] Train net output #141: loss3/loss17 = 0.84809 (* 0.0909091 = 0.0770991 loss) | |
| I0510 16:05:39.429137 10926 solver.cpp:245] Train net output #142: loss3/loss18 = 0.968665 (* 0.0909091 = 0.0880605 loss) | |
| I0510 16:05:39.429154 10926 solver.cpp:245] Train net output #143: loss3/loss19 = 1.15759 (* 0.0909091 = 0.105236 loss) | |
| I0510 16:05:39.429169 10926 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00753131 (* 0.0909091 = 0.000684664 loss) | |
| I0510 16:05:39.429183 10926 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00234936 (* 0.0909091 = 0.000213578 loss) | |
| I0510 16:05:39.429198 10926 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00352691 (* 0.0909091 = 0.000320628 loss) | |
| I0510 16:05:39.429211 10926 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0510 16:05:39.429224 10926 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0510 16:05:39.429240 10926 solver.cpp:245] Train net output #149: total_confidence = 5.18112e-05 | |
| I0510 16:05:39.429266 10926 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.000444044 | |
| I0510 16:05:39.429281 10926 sgd_solver.cpp:106] Iteration 8000, lr = 0.001 | |
| I0510 16:06:19.739804 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 32.6045 > 30) by scale factor 0.920118 | |
| I0510 16:06:41.144783 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 33.9991 > 30) by scale factor 0.882377 | |
| I0510 16:06:53.484930 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 43.4614 > 30) by scale factor 0.690267 | |
| I0510 16:07:13.982434 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 41.1215 > 30) by scale factor 0.729545 | |
| I0510 16:07:18.992120 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 31.9122 > 30) by scale factor 0.94008 | |
| I0510 16:07:22.504425 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 85.8835 > 30) by scale factor 0.34931 | |
| I0510 16:07:44.232645 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 30.3836 > 30) by scale factor 0.987375 | |
| I0510 16:07:46.287421 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 39.791 > 30) by scale factor 0.753939 | |
| I0510 16:08:06.196308 10926 solver.cpp:229] Iteration 8500, loss = 10.8464 | |
| I0510 16:08:06.196372 10926 solver.cpp:245] Train net output #0: loss1/accuracy = 0.025 | |
| I0510 16:08:06.196390 10926 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.125 | |
| I0510 16:08:06.196403 10926 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0510 16:08:06.196416 10926 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0510 16:08:06.196429 10926 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.125 | |
| I0510 16:08:06.196440 10926 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.5 | |
| I0510 16:08:06.196452 10926 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.5 | |
| I0510 16:08:06.196465 10926 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.875 | |
| I0510 16:08:06.196477 10926 solver.cpp:245] Train net output #8: loss1/accuracy08 = 1 | |
| I0510 16:08:06.196488 10926 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0510 16:08:06.196501 10926 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0510 16:08:06.196512 10926 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0510 16:08:06.196523 10926 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0510 16:08:06.196535 10926 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0510 16:08:06.196552 10926 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0510 16:08:06.196563 10926 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0510 16:08:06.196575 10926 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0510 16:08:06.196588 10926 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0510 16:08:06.196599 10926 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0510 16:08:06.196610 10926 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0510 16:08:06.196622 10926 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0510 16:08:06.196635 10926 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0510 16:08:06.196648 10926 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0510 16:08:06.196668 10926 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.772727 | |
| I0510 16:08:06.196682 10926 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.15 | |
| I0510 16:08:06.196698 10926 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.50952 (* 0.3 = 1.05285 loss) | |
| I0510 16:08:06.196713 10926 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 0.911243 (* 0.3 = 0.273373 loss) | |
| I0510 16:08:06.196728 10926 solver.cpp:245] Train net output #27: loss1/loss01 = 3.41175 (* 0.0272727 = 0.0930477 loss) | |
| I0510 16:08:06.196743 10926 solver.cpp:245] Train net output #28: loss1/loss02 = 3.47193 (* 0.0272727 = 0.0946889 loss) | |
| I0510 16:08:06.196758 10926 solver.cpp:245] Train net output #29: loss1/loss03 = 4.09948 (* 0.0272727 = 0.111804 loss) | |
| I0510 16:08:06.196771 10926 solver.cpp:245] Train net output #30: loss1/loss04 = 3.26689 (* 0.0272727 = 0.0890969 loss) | |
| I0510 16:08:06.196784 10926 solver.cpp:245] Train net output #31: loss1/loss05 = 2.49539 (* 0.0272727 = 0.068056 loss) | |
| I0510 16:08:06.196799 10926 solver.cpp:245] Train net output #32: loss1/loss06 = 2.57443 (* 0.0272727 = 0.0702117 loss) | |
| I0510 16:08:06.196813 10926 solver.cpp:245] Train net output #33: loss1/loss07 = 0.958988 (* 0.0272727 = 0.0261542 loss) | |
| I0510 16:08:06.196827 10926 solver.cpp:245] Train net output #34: loss1/loss08 = 0.0383247 (* 0.0272727 = 0.00104522 loss) | |
| I0510 16:08:06.196842 10926 solver.cpp:245] Train net output #35: loss1/loss09 = 0.0118997 (* 0.0272727 = 0.000324536 loss) | |
| I0510 16:08:06.196856 10926 solver.cpp:245] Train net output #36: loss1/loss10 = 0.00816988 (* 0.0272727 = 0.000222815 loss) | |
| I0510 16:08:06.196871 10926 solver.cpp:245] Train net output #37: loss1/loss11 = 0.00510479 (* 0.0272727 = 0.000139222 loss) | |
| I0510 16:08:06.196909 10926 solver.cpp:245] Train net output #38: loss1/loss12 = 0.00602855 (* 0.0272727 = 0.000164415 loss) | |
| I0510 16:08:06.196925 10926 solver.cpp:245] Train net output #39: loss1/loss13 = 0.0050662 (* 0.0272727 = 0.000138169 loss) | |
| I0510 16:08:06.196939 10926 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0054044 (* 0.0272727 = 0.000147393 loss) | |
| I0510 16:08:06.196954 10926 solver.cpp:245] Train net output #41: loss1/loss15 = 0.00463692 (* 0.0272727 = 0.000126461 loss) | |
| I0510 16:08:06.196967 10926 solver.cpp:245] Train net output #42: loss1/loss16 = 0.00222754 (* 0.0272727 = 6.07511e-05 loss) | |
| I0510 16:08:06.196981 10926 solver.cpp:245] Train net output #43: loss1/loss17 = 0.00387623 (* 0.0272727 = 0.000105715 loss) | |
| I0510 16:08:06.196995 10926 solver.cpp:245] Train net output #44: loss1/loss18 = 0.00312736 (* 0.0272727 = 8.52917e-05 loss) | |
| I0510 16:08:06.197010 10926 solver.cpp:245] Train net output #45: loss1/loss19 = 0.00228259 (* 0.0272727 = 6.22524e-05 loss) | |
| I0510 16:08:06.197024 10926 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00202615 (* 0.0272727 = 5.52587e-05 loss) | |
| I0510 16:08:06.197039 10926 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0019811 (* 0.0272727 = 5.403e-05 loss) | |
| I0510 16:08:06.197053 10926 solver.cpp:245] Train net output #48: loss1/loss22 = 0.000941349 (* 0.0272727 = 2.56732e-05 loss) | |
| I0510 16:08:06.197067 10926 solver.cpp:245] Train net output #49: loss2/accuracy = 0.025 | |
| I0510 16:08:06.197078 10926 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0510 16:08:06.197090 10926 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0510 16:08:06.197101 10926 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0.125 | |
| I0510 16:08:06.197113 10926 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.125 | |
| I0510 16:08:06.197139 10926 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.5 | |
| I0510 16:08:06.197149 10926 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.5 | |
| I0510 16:08:06.197157 10926 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.875 | |
| I0510 16:08:06.197170 10926 solver.cpp:245] Train net output #57: loss2/accuracy08 = 1 | |
| I0510 16:08:06.197181 10926 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0510 16:08:06.197198 10926 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0510 16:08:06.197221 10926 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0510 16:08:06.197243 10926 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0510 16:08:06.197262 10926 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0510 16:08:06.197274 10926 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0510 16:08:06.197286 10926 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0510 16:08:06.197299 10926 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0510 16:08:06.197309 10926 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0510 16:08:06.197320 10926 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0510 16:08:06.197332 10926 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0510 16:08:06.197345 10926 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0510 16:08:06.197355 10926 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0510 16:08:06.197367 10926 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0510 16:08:06.197378 10926 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.761364 | |
| I0510 16:08:06.197391 10926 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.15 | |
| I0510 16:08:06.197405 10926 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.57495 (* 0.3 = 1.07249 loss) | |
| I0510 16:08:06.197419 10926 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 0.943564 (* 0.3 = 0.283069 loss) | |
| I0510 16:08:06.197433 10926 solver.cpp:245] Train net output #76: loss2/loss01 = 3.28871 (* 0.0272727 = 0.0896922 loss) | |
| I0510 16:08:06.197463 10926 solver.cpp:245] Train net output #77: loss2/loss02 = 3.91019 (* 0.0272727 = 0.106641 loss) | |
| I0510 16:08:06.197479 10926 solver.cpp:245] Train net output #78: loss2/loss03 = 3.62985 (* 0.0272727 = 0.0989958 loss) | |
| I0510 16:08:06.197492 10926 solver.cpp:245] Train net output #79: loss2/loss04 = 3.81341 (* 0.0272727 = 0.104002 loss) | |
| I0510 16:08:06.197506 10926 solver.cpp:245] Train net output #80: loss2/loss05 = 1.69632 (* 0.0272727 = 0.0462632 loss) | |
| I0510 16:08:06.197520 10926 solver.cpp:245] Train net output #81: loss2/loss06 = 2.29652 (* 0.0272727 = 0.0626324 loss) | |
| I0510 16:08:06.197535 10926 solver.cpp:245] Train net output #82: loss2/loss07 = 0.923241 (* 0.0272727 = 0.0251793 loss) | |
| I0510 16:08:06.197548 10926 solver.cpp:245] Train net output #83: loss2/loss08 = 0.0421195 (* 0.0272727 = 0.00114871 loss) | |
| I0510 16:08:06.197563 10926 solver.cpp:245] Train net output #84: loss2/loss09 = 0.0159685 (* 0.0272727 = 0.000435506 loss) | |
| I0510 16:08:06.197577 10926 solver.cpp:245] Train net output #85: loss2/loss10 = 0.00914905 (* 0.0272727 = 0.000249519 loss) | |
| I0510 16:08:06.197594 10926 solver.cpp:245] Train net output #86: loss2/loss11 = 0.00677653 (* 0.0272727 = 0.000184815 loss) | |
| I0510 16:08:06.197609 10926 solver.cpp:245] Train net output #87: loss2/loss12 = 0.00618418 (* 0.0272727 = 0.00016866 loss) | |
| I0510 16:08:06.197624 10926 solver.cpp:245] Train net output #88: loss2/loss13 = 0.00533965 (* 0.0272727 = 0.000145627 loss) | |
| I0510 16:08:06.197638 10926 solver.cpp:245] Train net output #89: loss2/loss14 = 0.00521109 (* 0.0272727 = 0.000142121 loss) | |
| I0510 16:08:06.197652 10926 solver.cpp:245] Train net output #90: loss2/loss15 = 0.00512735 (* 0.0272727 = 0.000139837 loss) | |
| I0510 16:08:06.197666 10926 solver.cpp:245] Train net output #91: loss2/loss16 = 0.00333289 (* 0.0272727 = 9.0897e-05 loss) | |
| I0510 16:08:06.197681 10926 solver.cpp:245] Train net output #92: loss2/loss17 = 0.00186607 (* 0.0272727 = 5.08927e-05 loss) | |
| I0510 16:08:06.197697 10926 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0012419 (* 0.0272727 = 3.38699e-05 loss) | |
| I0510 16:08:06.197712 10926 solver.cpp:245] Train net output #94: loss2/loss19 = 0.00235182 (* 0.0272727 = 6.41406e-05 loss) | |
| I0510 16:08:06.197727 10926 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00224313 (* 0.0272727 = 6.11764e-05 loss) | |
| I0510 16:08:06.197741 10926 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00153764 (* 0.0272727 = 4.19356e-05 loss) | |
| I0510 16:08:06.197756 10926 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00119461 (* 0.0272727 = 3.25803e-05 loss) | |
| I0510 16:08:06.197768 10926 solver.cpp:245] Train net output #98: loss3/accuracy = 0.075 | |
| I0510 16:08:06.197780 10926 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0510 16:08:06.197793 10926 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.125 | |
| I0510 16:08:06.197803 10926 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0510 16:08:06.197815 10926 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.25 | |
| I0510 16:08:06.197827 10926 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.375 | |
| I0510 16:08:06.197839 10926 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.5 | |
| I0510 16:08:06.197850 10926 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.875 | |
| I0510 16:08:06.197861 10926 solver.cpp:245] Train net output #106: loss3/accuracy08 = 1 | |
| I0510 16:08:06.197880 10926 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0510 16:08:06.197904 10926 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0510 16:08:06.197919 10926 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0510 16:08:06.197932 10926 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0510 16:08:06.197942 10926 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0510 16:08:06.197954 10926 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0510 16:08:06.197978 10926 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0510 16:08:06.197991 10926 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0510 16:08:06.198002 10926 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0510 16:08:06.198014 10926 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0510 16:08:06.198025 10926 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0510 16:08:06.198037 10926 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0510 16:08:06.198050 10926 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0510 16:08:06.198060 10926 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0510 16:08:06.198072 10926 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.778409 | |
| I0510 16:08:06.198084 10926 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.175 | |
| I0510 16:08:06.198098 10926 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.38545 (* 1 = 3.38545 loss) | |
| I0510 16:08:06.198112 10926 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 0.888198 (* 1 = 0.888198 loss) | |
| I0510 16:08:06.198127 10926 solver.cpp:245] Train net output #125: loss3/loss01 = 2.99598 (* 0.0909091 = 0.272362 loss) | |
| I0510 16:08:06.198140 10926 solver.cpp:245] Train net output #126: loss3/loss02 = 3.35823 (* 0.0909091 = 0.305294 loss) | |
| I0510 16:08:06.198154 10926 solver.cpp:245] Train net output #127: loss3/loss03 = 3.42418 (* 0.0909091 = 0.311289 loss) | |
| I0510 16:08:06.198168 10926 solver.cpp:245] Train net output #128: loss3/loss04 = 3.28982 (* 0.0909091 = 0.299075 loss) | |
| I0510 16:08:06.198182 10926 solver.cpp:245] Train net output #129: loss3/loss05 = 2.09115 (* 0.0909091 = 0.190104 loss) | |
| I0510 16:08:06.198197 10926 solver.cpp:245] Train net output #130: loss3/loss06 = 2.38755 (* 0.0909091 = 0.21705 loss) | |
| I0510 16:08:06.198211 10926 solver.cpp:245] Train net output #131: loss3/loss07 = 0.761127 (* 0.0909091 = 0.0691933 loss) | |
| I0510 16:08:06.198225 10926 solver.cpp:245] Train net output #132: loss3/loss08 = 0.0499319 (* 0.0909091 = 0.00453926 loss) | |
| I0510 16:08:06.198238 10926 solver.cpp:245] Train net output #133: loss3/loss09 = 0.00287784 (* 0.0909091 = 0.000261622 loss) | |
| I0510 16:08:06.198252 10926 solver.cpp:245] Train net output #134: loss3/loss10 = 0.00292598 (* 0.0909091 = 0.000265998 loss) | |
| I0510 16:08:06.198266 10926 solver.cpp:245] Train net output #135: loss3/loss11 = 0.00197346 (* 0.0909091 = 0.000179406 loss) | |
| I0510 16:08:06.198281 10926 solver.cpp:245] Train net output #136: loss3/loss12 = 0.00102671 (* 0.0909091 = 9.33371e-05 loss) | |
| I0510 16:08:06.198294 10926 solver.cpp:245] Train net output #137: loss3/loss13 = 0.00131627 (* 0.0909091 = 0.000119661 loss) | |
| I0510 16:08:06.198308 10926 solver.cpp:245] Train net output #138: loss3/loss14 = 0.000965934 (* 0.0909091 = 8.78122e-05 loss) | |
| I0510 16:08:06.198323 10926 solver.cpp:245] Train net output #139: loss3/loss15 = 0.000752404 (* 0.0909091 = 6.84003e-05 loss) | |
| I0510 16:08:06.198336 10926 solver.cpp:245] Train net output #140: loss3/loss16 = 0.000531871 (* 0.0909091 = 4.8352e-05 loss) | |
| I0510 16:08:06.198350 10926 solver.cpp:245] Train net output #141: loss3/loss17 = 0.000628605 (* 0.0909091 = 5.71459e-05 loss) | |
| I0510 16:08:06.198364 10926 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00075822 (* 0.0909091 = 6.89291e-05 loss) | |
| I0510 16:08:06.198379 10926 solver.cpp:245] Train net output #143: loss3/loss19 = 0.000578433 (* 0.0909091 = 5.25849e-05 loss) | |
| I0510 16:08:06.198392 10926 solver.cpp:245] Train net output #144: loss3/loss20 = 0.000542828 (* 0.0909091 = 4.9348e-05 loss) | |
| I0510 16:08:06.198406 10926 solver.cpp:245] Train net output #145: loss3/loss21 = 0.000476946 (* 0.0909091 = 4.33588e-05 loss) | |
| I0510 16:08:06.198421 10926 solver.cpp:245] Train net output #146: loss3/loss22 = 0.000325442 (* 0.0909091 = 2.95856e-05 loss) | |
| I0510 16:08:06.198442 10926 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0510 16:08:06.198456 10926 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0510 16:08:06.198467 10926 solver.cpp:245] Train net output #149: total_confidence = 2.94861e-05 | |
| I0510 16:08:06.198479 10926 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.000443809 | |
| I0510 16:08:06.198493 10926 sgd_solver.cpp:106] Iteration 8500, lr = 0.001 | |
| I0510 16:09:14.323185 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 57.9867 > 30) by scale factor 0.51736 | |
| I0510 16:09:46.619323 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 43.9038 > 30) by scale factor 0.683312 | |
| I0510 16:09:53.965337 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 30.2336 > 30) by scale factor 0.992272 | |
| I0510 16:10:32.943763 10926 solver.cpp:229] Iteration 9000, loss = 10.6964 | |
| I0510 16:10:32.943940 10926 solver.cpp:245] Train net output #0: loss1/accuracy = 0 | |
| I0510 16:10:32.943961 10926 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0510 16:10:32.943974 10926 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.125 | |
| I0510 16:10:32.943989 10926 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0510 16:10:32.944000 10926 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0 | |
| I0510 16:10:32.944012 10926 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.375 | |
| I0510 16:10:32.944025 10926 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.625 | |
| I0510 16:10:32.944037 10926 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.75 | |
| I0510 16:10:32.944051 10926 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.875 | |
| I0510 16:10:32.944062 10926 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0510 16:10:32.944074 10926 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0510 16:10:32.944087 10926 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0510 16:10:32.944098 10926 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0510 16:10:32.944110 10926 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0510 16:10:32.944123 10926 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0510 16:10:32.944134 10926 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0510 16:10:32.944146 10926 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0510 16:10:32.944159 10926 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0510 16:10:32.944170 10926 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0510 16:10:32.944183 10926 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0510 16:10:32.944195 10926 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0510 16:10:32.944207 10926 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0510 16:10:32.944219 10926 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0510 16:10:32.944231 10926 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.738636 | |
| I0510 16:10:32.944243 10926 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.170732 | |
| I0510 16:10:32.944260 10926 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.48471 (* 0.3 = 1.04541 loss) | |
| I0510 16:10:32.944275 10926 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.07545 (* 0.3 = 0.322636 loss) | |
| I0510 16:10:32.944291 10926 solver.cpp:245] Train net output #27: loss1/loss01 = 3.71971 (* 0.0272727 = 0.101447 loss) | |
| I0510 16:10:32.944305 10926 solver.cpp:245] Train net output #28: loss1/loss02 = 3.84913 (* 0.0272727 = 0.104976 loss) | |
| I0510 16:10:32.944319 10926 solver.cpp:245] Train net output #29: loss1/loss03 = 3.7972 (* 0.0272727 = 0.10356 loss) | |
| I0510 16:10:32.944334 10926 solver.cpp:245] Train net output #30: loss1/loss04 = 4.37062 (* 0.0272727 = 0.119199 loss) | |
| I0510 16:10:32.944349 10926 solver.cpp:245] Train net output #31: loss1/loss05 = 3.14171 (* 0.0272727 = 0.0856831 loss) | |
| I0510 16:10:32.944362 10926 solver.cpp:245] Train net output #32: loss1/loss06 = 2.14545 (* 0.0272727 = 0.0585123 loss) | |
| I0510 16:10:32.944376 10926 solver.cpp:245] Train net output #33: loss1/loss07 = 1.65786 (* 0.0272727 = 0.0452142 loss) | |
| I0510 16:10:32.944391 10926 solver.cpp:245] Train net output #34: loss1/loss08 = 0.786938 (* 0.0272727 = 0.0214619 loss) | |
| I0510 16:10:32.944406 10926 solver.cpp:245] Train net output #35: loss1/loss09 = 0.0797418 (* 0.0272727 = 0.00217478 loss) | |
| I0510 16:10:32.944421 10926 solver.cpp:245] Train net output #36: loss1/loss10 = 0.0865031 (* 0.0272727 = 0.00235918 loss) | |
| I0510 16:10:32.944435 10926 solver.cpp:245] Train net output #37: loss1/loss11 = 0.0622936 (* 0.0272727 = 0.00169892 loss) | |
| I0510 16:10:32.944449 10926 solver.cpp:245] Train net output #38: loss1/loss12 = 0.0358353 (* 0.0272727 = 0.000977328 loss) | |
| I0510 16:10:32.944464 10926 solver.cpp:245] Train net output #39: loss1/loss13 = 0.0518409 (* 0.0272727 = 0.00141384 loss) | |
| I0510 16:10:32.944501 10926 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0431674 (* 0.0272727 = 0.00117729 loss) | |
| I0510 16:10:32.944517 10926 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0322645 (* 0.0272727 = 0.00087994 loss) | |
| I0510 16:10:32.944532 10926 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0332591 (* 0.0272727 = 0.000907066 loss) | |
| I0510 16:10:32.944546 10926 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0282482 (* 0.0272727 = 0.000770406 loss) | |
| I0510 16:10:32.944561 10926 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0298163 (* 0.0272727 = 0.000813173 loss) | |
| I0510 16:10:32.944576 10926 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0145963 (* 0.0272727 = 0.000398081 loss) | |
| I0510 16:10:32.944589 10926 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0220997 (* 0.0272727 = 0.000602719 loss) | |
| I0510 16:10:32.944604 10926 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0156009 (* 0.0272727 = 0.000425479 loss) | |
| I0510 16:10:32.944618 10926 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0199011 (* 0.0272727 = 0.000542758 loss) | |
| I0510 16:10:32.944631 10926 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0243902 | |
| I0510 16:10:32.944644 10926 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.375 | |
| I0510 16:10:32.944656 10926 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0.125 | |
| I0510 16:10:32.944669 10926 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0510 16:10:32.944680 10926 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0 | |
| I0510 16:10:32.944692 10926 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.5 | |
| I0510 16:10:32.944705 10926 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.5 | |
| I0510 16:10:32.944716 10926 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.75 | |
| I0510 16:10:32.944728 10926 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.875 | |
| I0510 16:10:32.944741 10926 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0510 16:10:32.944752 10926 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0510 16:10:32.944766 10926 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0510 16:10:32.944777 10926 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0510 16:10:32.944789 10926 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0510 16:10:32.944800 10926 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0510 16:10:32.944813 10926 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0510 16:10:32.944824 10926 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0510 16:10:32.944836 10926 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0510 16:10:32.944849 10926 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0510 16:10:32.944859 10926 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0510 16:10:32.944871 10926 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0510 16:10:32.944887 10926 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0510 16:10:32.944900 10926 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0510 16:10:32.944911 10926 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.744318 | |
| I0510 16:10:32.944922 10926 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.146341 | |
| I0510 16:10:32.944936 10926 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.53969 (* 0.3 = 1.06191 loss) | |
| I0510 16:10:32.944957 10926 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.03175 (* 0.3 = 0.309526 loss) | |
| I0510 16:10:32.944972 10926 solver.cpp:245] Train net output #76: loss2/loss01 = 3.09175 (* 0.0272727 = 0.0843205 loss) | |
| I0510 16:10:32.944986 10926 solver.cpp:245] Train net output #77: loss2/loss02 = 3.23193 (* 0.0272727 = 0.0881436 loss) | |
| I0510 16:10:32.945000 10926 solver.cpp:245] Train net output #78: loss2/loss03 = 3.92868 (* 0.0272727 = 0.107146 loss) | |
| I0510 16:10:32.945026 10926 solver.cpp:245] Train net output #79: loss2/loss04 = 4.03313 (* 0.0272727 = 0.109994 loss) | |
| I0510 16:10:32.945041 10926 solver.cpp:245] Train net output #80: loss2/loss05 = 2.78111 (* 0.0272727 = 0.0758484 loss) | |
| I0510 16:10:32.945056 10926 solver.cpp:245] Train net output #81: loss2/loss06 = 2.00717 (* 0.0272727 = 0.054741 loss) | |
| I0510 16:10:32.945070 10926 solver.cpp:245] Train net output #82: loss2/loss07 = 1.37127 (* 0.0272727 = 0.0373984 loss) | |
| I0510 16:10:32.945085 10926 solver.cpp:245] Train net output #83: loss2/loss08 = 0.774877 (* 0.0272727 = 0.021133 loss) | |
| I0510 16:10:32.945098 10926 solver.cpp:245] Train net output #84: loss2/loss09 = 0.0688882 (* 0.0272727 = 0.00187877 loss) | |
| I0510 16:10:32.945113 10926 solver.cpp:245] Train net output #85: loss2/loss10 = 0.0303917 (* 0.0272727 = 0.000828864 loss) | |
| I0510 16:10:32.945145 10926 solver.cpp:245] Train net output #86: loss2/loss11 = 0.0261605 (* 0.0272727 = 0.000713468 loss) | |
| I0510 16:10:32.945161 10926 solver.cpp:245] Train net output #87: loss2/loss12 = 0.0300773 (* 0.0272727 = 0.000820289 loss) | |
| I0510 16:10:32.945175 10926 solver.cpp:245] Train net output #88: loss2/loss13 = 0.0235631 (* 0.0272727 = 0.000642629 loss) | |
| I0510 16:10:32.945190 10926 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0166495 (* 0.0272727 = 0.000454077 loss) | |
| I0510 16:10:32.945204 10926 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0133167 (* 0.0272727 = 0.000363182 loss) | |
| I0510 16:10:32.945219 10926 solver.cpp:245] Train net output #91: loss2/loss16 = 0.022784 (* 0.0272727 = 0.000621381 loss) | |
| I0510 16:10:32.945233 10926 solver.cpp:245] Train net output #92: loss2/loss17 = 0.00707186 (* 0.0272727 = 0.000192869 loss) | |
| I0510 16:10:32.945247 10926 solver.cpp:245] Train net output #93: loss2/loss18 = 0.00764761 (* 0.0272727 = 0.000208571 loss) | |
| I0510 16:10:32.945261 10926 solver.cpp:245] Train net output #94: loss2/loss19 = 0.00888341 (* 0.0272727 = 0.000242275 loss) | |
| I0510 16:10:32.945276 10926 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00515869 (* 0.0272727 = 0.000140692 loss) | |
| I0510 16:10:32.945291 10926 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00491042 (* 0.0272727 = 0.00013392 loss) | |
| I0510 16:10:32.945304 10926 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00757126 (* 0.0272727 = 0.000206489 loss) | |
| I0510 16:10:32.945317 10926 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0731707 | |
| I0510 16:10:32.945329 10926 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.375 | |
| I0510 16:10:32.945341 10926 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.125 | |
| I0510 16:10:32.945353 10926 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0510 16:10:32.945364 10926 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0 | |
| I0510 16:10:32.945376 10926 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.25 | |
| I0510 16:10:32.945389 10926 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.625 | |
| I0510 16:10:32.945400 10926 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.75 | |
| I0510 16:10:32.945412 10926 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.875 | |
| I0510 16:10:32.945423 10926 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0510 16:10:32.945436 10926 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0510 16:10:32.945447 10926 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0510 16:10:32.945459 10926 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0510 16:10:32.945472 10926 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0510 16:10:32.945484 10926 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0510 16:10:32.945495 10926 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0510 16:10:32.945508 10926 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0510 16:10:32.945531 10926 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0510 16:10:32.945544 10926 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0510 16:10:32.945556 10926 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0510 16:10:32.945569 10926 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0510 16:10:32.945580 10926 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0510 16:10:32.945592 10926 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0510 16:10:32.945605 10926 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.755682 | |
| I0510 16:10:32.945616 10926 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.146341 | |
| I0510 16:10:32.945631 10926 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.36481 (* 1 = 3.36481 loss) | |
| I0510 16:10:32.945644 10926 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 0.973553 (* 1 = 0.973553 loss) | |
| I0510 16:10:32.945658 10926 solver.cpp:245] Train net output #125: loss3/loss01 = 2.72533 (* 0.0909091 = 0.247757 loss) | |
| I0510 16:10:32.945673 10926 solver.cpp:245] Train net output #126: loss3/loss02 = 3.33989 (* 0.0909091 = 0.303626 loss) | |
| I0510 16:10:32.945688 10926 solver.cpp:245] Train net output #127: loss3/loss03 = 3.86874 (* 0.0909091 = 0.351704 loss) | |
| I0510 16:10:32.945701 10926 solver.cpp:245] Train net output #128: loss3/loss04 = 3.86013 (* 0.0909091 = 0.35092 loss) | |
| I0510 16:10:32.945715 10926 solver.cpp:245] Train net output #129: loss3/loss05 = 2.81643 (* 0.0909091 = 0.256039 loss) | |
| I0510 16:10:32.945729 10926 solver.cpp:245] Train net output #130: loss3/loss06 = 1.87351 (* 0.0909091 = 0.170319 loss) | |
| I0510 16:10:32.945744 10926 solver.cpp:245] Train net output #131: loss3/loss07 = 1.29311 (* 0.0909091 = 0.117555 loss) | |
| I0510 16:10:32.945757 10926 solver.cpp:245] Train net output #132: loss3/loss08 = 0.61915 (* 0.0909091 = 0.0562864 loss) | |
| I0510 16:10:32.945772 10926 solver.cpp:245] Train net output #133: loss3/loss09 = 0.0338372 (* 0.0909091 = 0.00307611 loss) | |
| I0510 16:10:32.945786 10926 solver.cpp:245] Train net output #134: loss3/loss10 = 0.021464 (* 0.0909091 = 0.00195127 loss) | |
| I0510 16:10:32.945801 10926 solver.cpp:245] Train net output #135: loss3/loss11 = 0.00669557 (* 0.0909091 = 0.000608688 loss) | |
| I0510 16:10:32.945814 10926 solver.cpp:245] Train net output #136: loss3/loss12 = 0.00817836 (* 0.0909091 = 0.000743487 loss) | |
| I0510 16:10:32.945828 10926 solver.cpp:245] Train net output #137: loss3/loss13 = 0.00879542 (* 0.0909091 = 0.000799584 loss) | |
| I0510 16:10:32.945842 10926 solver.cpp:245] Train net output #138: loss3/loss14 = 0.014926 (* 0.0909091 = 0.00135691 loss) | |
| I0510 16:10:32.945858 10926 solver.cpp:245] Train net output #139: loss3/loss15 = 0.00773003 (* 0.0909091 = 0.00070273 loss) | |
| I0510 16:10:32.945871 10926 solver.cpp:245] Train net output #140: loss3/loss16 = 0.0069032 (* 0.0909091 = 0.000627564 loss) | |
| I0510 16:10:32.945885 10926 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00579381 (* 0.0909091 = 0.00052671 loss) | |
| I0510 16:10:32.945899 10926 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00387328 (* 0.0909091 = 0.000352117 loss) | |
| I0510 16:10:32.945914 10926 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00492363 (* 0.0909091 = 0.000447603 loss) | |
| I0510 16:10:32.945930 10926 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00224923 (* 0.0909091 = 0.000204475 loss) | |
| I0510 16:10:32.945945 10926 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00271907 (* 0.0909091 = 0.000247188 loss) | |
| I0510 16:10:32.945960 10926 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00179884 (* 0.0909091 = 0.000163531 loss) | |
| I0510 16:10:32.945972 10926 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0510 16:10:32.945986 10926 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0510 16:10:32.946001 10926 solver.cpp:245] Train net output #149: total_confidence = 3.94355e-06 | |
| I0510 16:10:32.946023 10926 solver.cpp:245] Train net output #150: total_confidence_not_rec = 2.5612e-05 | |
| I0510 16:10:32.946038 10926 sgd_solver.cpp:106] Iteration 9000, lr = 0.001 | |
| I0510 16:10:49.004371 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 31.5447 > 30) by scale factor 0.951033 | |
| I0510 16:11:02.550953 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 33.983 > 30) by scale factor 0.882795 | |
| I0510 16:11:29.277021 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 74.5215 > 30) by scale factor 0.402568 | |
| I0510 16:12:34.964128 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 32.5793 > 30) by scale factor 0.920829 | |
| I0510 16:12:52.272101 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 38.4121 > 30) by scale factor 0.781003 | |
| I0510 16:12:59.799396 10926 solver.cpp:229] Iteration 9500, loss = 10.6149 | |
| I0510 16:12:59.799466 10926 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0909091 | |
| I0510 16:12:59.799485 10926 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0510 16:12:59.799499 10926 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0510 16:12:59.799512 10926 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0510 16:12:59.799525 10926 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.125 | |
| I0510 16:12:59.799536 10926 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.5 | |
| I0510 16:12:59.799549 10926 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.5 | |
| I0510 16:12:59.799561 10926 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.625 | |
| I0510 16:12:59.799573 10926 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.875 | |
| I0510 16:12:59.799587 10926 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0510 16:12:59.799602 10926 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0510 16:12:59.799614 10926 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0510 16:12:59.799626 10926 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0510 16:12:59.799639 10926 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0510 16:12:59.799651 10926 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0510 16:12:59.799664 10926 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0510 16:12:59.799677 10926 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0510 16:12:59.799690 10926 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0510 16:12:59.799702 10926 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0510 16:12:59.799715 10926 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0510 16:12:59.799726 10926 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0510 16:12:59.799738 10926 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0510 16:12:59.799751 10926 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0510 16:12:59.799762 10926 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.767045 | |
| I0510 16:12:59.799774 10926 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.227273 | |
| I0510 16:12:59.799792 10926 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.39294 (* 0.3 = 1.01788 loss) | |
| I0510 16:12:59.799805 10926 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 0.956352 (* 0.3 = 0.286906 loss) | |
| I0510 16:12:59.799820 10926 solver.cpp:245] Train net output #27: loss1/loss01 = 3.66113 (* 0.0272727 = 0.0998491 loss) | |
| I0510 16:12:59.799835 10926 solver.cpp:245] Train net output #28: loss1/loss02 = 3.84117 (* 0.0272727 = 0.104759 loss) | |
| I0510 16:12:59.799849 10926 solver.cpp:245] Train net output #29: loss1/loss03 = 3.49596 (* 0.0272727 = 0.0953444 loss) | |
| I0510 16:12:59.799863 10926 solver.cpp:245] Train net output #30: loss1/loss04 = 3.44299 (* 0.0272727 = 0.0938996 loss) | |
| I0510 16:12:59.799877 10926 solver.cpp:245] Train net output #31: loss1/loss05 = 2.58916 (* 0.0272727 = 0.0706134 loss) | |
| I0510 16:12:59.799891 10926 solver.cpp:245] Train net output #32: loss1/loss06 = 2.47042 (* 0.0272727 = 0.0673752 loss) | |
| I0510 16:12:59.799906 10926 solver.cpp:245] Train net output #33: loss1/loss07 = 2.10519 (* 0.0272727 = 0.0574142 loss) | |
| I0510 16:12:59.799919 10926 solver.cpp:245] Train net output #34: loss1/loss08 = 0.796668 (* 0.0272727 = 0.0217273 loss) | |
| I0510 16:12:59.799934 10926 solver.cpp:245] Train net output #35: loss1/loss09 = 0.014891 (* 0.0272727 = 0.000406117 loss) | |
| I0510 16:12:59.799949 10926 solver.cpp:245] Train net output #36: loss1/loss10 = 0.00969185 (* 0.0272727 = 0.000264323 loss) | |
| I0510 16:12:59.799963 10926 solver.cpp:245] Train net output #37: loss1/loss11 = 0.011731 (* 0.0272727 = 0.000319937 loss) | |
| I0510 16:12:59.800012 10926 solver.cpp:245] Train net output #38: loss1/loss12 = 0.00849548 (* 0.0272727 = 0.000231695 loss) | |
| I0510 16:12:59.800029 10926 solver.cpp:245] Train net output #39: loss1/loss13 = 0.00629522 (* 0.0272727 = 0.000171688 loss) | |
| I0510 16:12:59.800043 10926 solver.cpp:245] Train net output #40: loss1/loss14 = 0.00694127 (* 0.0272727 = 0.000189307 loss) | |
| I0510 16:12:59.800057 10926 solver.cpp:245] Train net output #41: loss1/loss15 = 0.00784169 (* 0.0272727 = 0.000213864 loss) | |
| I0510 16:12:59.800071 10926 solver.cpp:245] Train net output #42: loss1/loss16 = 0.00402673 (* 0.0272727 = 0.00010982 loss) | |
| I0510 16:12:59.800086 10926 solver.cpp:245] Train net output #43: loss1/loss17 = 0.00267081 (* 0.0272727 = 7.28402e-05 loss) | |
| I0510 16:12:59.800101 10926 solver.cpp:245] Train net output #44: loss1/loss18 = 0.00447546 (* 0.0272727 = 0.000122058 loss) | |
| I0510 16:12:59.800114 10926 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0035532 (* 0.0272727 = 9.69056e-05 loss) | |
| I0510 16:12:59.800128 10926 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0035271 (* 0.0272727 = 9.61937e-05 loss) | |
| I0510 16:12:59.800143 10926 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00158321 (* 0.0272727 = 4.31785e-05 loss) | |
| I0510 16:12:59.800158 10926 solver.cpp:245] Train net output #48: loss1/loss22 = 0.00213609 (* 0.0272727 = 5.82571e-05 loss) | |
| I0510 16:12:59.800169 10926 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0227273 | |
| I0510 16:12:59.800182 10926 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0510 16:12:59.800194 10926 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0510 16:12:59.800206 10926 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0510 16:12:59.800217 10926 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.125 | |
| I0510 16:12:59.800230 10926 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.375 | |
| I0510 16:12:59.800242 10926 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.375 | |
| I0510 16:12:59.800251 10926 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.625 | |
| I0510 16:12:59.800261 10926 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.875 | |
| I0510 16:12:59.800273 10926 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0510 16:12:59.800285 10926 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0510 16:12:59.800297 10926 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0510 16:12:59.800309 10926 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0510 16:12:59.800321 10926 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0510 16:12:59.800333 10926 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0510 16:12:59.800345 10926 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0510 16:12:59.800356 10926 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0510 16:12:59.800369 10926 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0510 16:12:59.800380 10926 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0510 16:12:59.800391 10926 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0510 16:12:59.800403 10926 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0510 16:12:59.800415 10926 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0510 16:12:59.800426 10926 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0510 16:12:59.800438 10926 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.755682 | |
| I0510 16:12:59.800451 10926 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.25 | |
| I0510 16:12:59.800464 10926 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.26402 (* 0.3 = 0.979206 loss) | |
| I0510 16:12:59.800478 10926 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 0.937847 (* 0.3 = 0.281354 loss) | |
| I0510 16:12:59.800493 10926 solver.cpp:245] Train net output #76: loss2/loss01 = 3.63632 (* 0.0272727 = 0.0991723 loss) | |
| I0510 16:12:59.800519 10926 solver.cpp:245] Train net output #77: loss2/loss02 = 4.00822 (* 0.0272727 = 0.109315 loss) | |
| I0510 16:12:59.800535 10926 solver.cpp:245] Train net output #78: loss2/loss03 = 4.05558 (* 0.0272727 = 0.110607 loss) | |
| I0510 16:12:59.800549 10926 solver.cpp:245] Train net output #79: loss2/loss04 = 3.41516 (* 0.0272727 = 0.0931408 loss) | |
| I0510 16:12:59.800564 10926 solver.cpp:245] Train net output #80: loss2/loss05 = 2.8666 (* 0.0272727 = 0.0781799 loss) | |
| I0510 16:12:59.800578 10926 solver.cpp:245] Train net output #81: loss2/loss06 = 2.29959 (* 0.0272727 = 0.0627161 loss) | |
| I0510 16:12:59.800592 10926 solver.cpp:245] Train net output #82: loss2/loss07 = 2.41043 (* 0.0272727 = 0.0657391 loss) | |
| I0510 16:12:59.800606 10926 solver.cpp:245] Train net output #83: loss2/loss08 = 0.984323 (* 0.0272727 = 0.0268452 loss) | |
| I0510 16:12:59.800621 10926 solver.cpp:245] Train net output #84: loss2/loss09 = 0.0073846 (* 0.0272727 = 0.000201398 loss) | |
| I0510 16:12:59.800635 10926 solver.cpp:245] Train net output #85: loss2/loss10 = 0.00625139 (* 0.0272727 = 0.000170492 loss) | |
| I0510 16:12:59.800653 10926 solver.cpp:245] Train net output #86: loss2/loss11 = 0.00592968 (* 0.0272727 = 0.000161719 loss) | |
| I0510 16:12:59.800668 10926 solver.cpp:245] Train net output #87: loss2/loss12 = 0.00578581 (* 0.0272727 = 0.000157795 loss) | |
| I0510 16:12:59.800683 10926 solver.cpp:245] Train net output #88: loss2/loss13 = 0.00369677 (* 0.0272727 = 0.000100821 loss) | |
| I0510 16:12:59.800696 10926 solver.cpp:245] Train net output #89: loss2/loss14 = 0.00430888 (* 0.0272727 = 0.000117515 loss) | |
| I0510 16:12:59.800711 10926 solver.cpp:245] Train net output #90: loss2/loss15 = 0.00407051 (* 0.0272727 = 0.000111014 loss) | |
| I0510 16:12:59.800727 10926 solver.cpp:245] Train net output #91: loss2/loss16 = 0.00367055 (* 0.0272727 = 0.000100106 loss) | |
| I0510 16:12:59.800742 10926 solver.cpp:245] Train net output #92: loss2/loss17 = 0.00280545 (* 0.0272727 = 7.65123e-05 loss) | |
| I0510 16:12:59.800756 10926 solver.cpp:245] Train net output #93: loss2/loss18 = 0.00144108 (* 0.0272727 = 3.93023e-05 loss) | |
| I0510 16:12:59.800771 10926 solver.cpp:245] Train net output #94: loss2/loss19 = 0.00389364 (* 0.0272727 = 0.00010619 loss) | |
| I0510 16:12:59.800786 10926 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00134787 (* 0.0272727 = 3.67601e-05 loss) | |
| I0510 16:12:59.800801 10926 solver.cpp:245] Train net output #96: loss2/loss21 = 0.0017527 (* 0.0272727 = 4.7801e-05 loss) | |
| I0510 16:12:59.800814 10926 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00101141 (* 0.0272727 = 2.7584e-05 loss) | |
| I0510 16:12:59.800827 10926 solver.cpp:245] Train net output #98: loss3/accuracy = 0.113636 | |
| I0510 16:12:59.800839 10926 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0510 16:12:59.800851 10926 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.125 | |
| I0510 16:12:59.800863 10926 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0510 16:12:59.800875 10926 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.25 | |
| I0510 16:12:59.800886 10926 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.375 | |
| I0510 16:12:59.800899 10926 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.5 | |
| I0510 16:12:59.800910 10926 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.625 | |
| I0510 16:12:59.800922 10926 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.875 | |
| I0510 16:12:59.800935 10926 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0510 16:12:59.800946 10926 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0510 16:12:59.800957 10926 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0510 16:12:59.800968 10926 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0510 16:12:59.800981 10926 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0510 16:12:59.800992 10926 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0510 16:12:59.801014 10926 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0510 16:12:59.801028 10926 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0510 16:12:59.801040 10926 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0510 16:12:59.801053 10926 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0510 16:12:59.801064 10926 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0510 16:12:59.801075 10926 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0510 16:12:59.801087 10926 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0510 16:12:59.801098 10926 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0510 16:12:59.801110 10926 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.778409 | |
| I0510 16:12:59.801137 10926 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.181818 | |
| I0510 16:12:59.801153 10926 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.15303 (* 1 = 3.15303 loss) | |
| I0510 16:12:59.801168 10926 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 0.850007 (* 1 = 0.850007 loss) | |
| I0510 16:12:59.801182 10926 solver.cpp:245] Train net output #125: loss3/loss01 = 3.11559 (* 0.0909091 = 0.283235 loss) | |
| I0510 16:12:59.801197 10926 solver.cpp:245] Train net output #126: loss3/loss02 = 3.29225 (* 0.0909091 = 0.299296 loss) | |
| I0510 16:12:59.801210 10926 solver.cpp:245] Train net output #127: loss3/loss03 = 3.60103 (* 0.0909091 = 0.327366 loss) | |
| I0510 16:12:59.801224 10926 solver.cpp:245] Train net output #128: loss3/loss04 = 2.66692 (* 0.0909091 = 0.242447 loss) | |
| I0510 16:12:59.801239 10926 solver.cpp:245] Train net output #129: loss3/loss05 = 2.26509 (* 0.0909091 = 0.205917 loss) | |
| I0510 16:12:59.801252 10926 solver.cpp:245] Train net output #130: loss3/loss06 = 1.80708 (* 0.0909091 = 0.16428 loss) | |
| I0510 16:12:59.801266 10926 solver.cpp:245] Train net output #131: loss3/loss07 = 1.5747 (* 0.0909091 = 0.143154 loss) | |
| I0510 16:12:59.801280 10926 solver.cpp:245] Train net output #132: loss3/loss08 = 0.572012 (* 0.0909091 = 0.0520011 loss) | |
| I0510 16:12:59.801295 10926 solver.cpp:245] Train net output #133: loss3/loss09 = 0.00192097 (* 0.0909091 = 0.000174633 loss) | |
| I0510 16:12:59.801309 10926 solver.cpp:245] Train net output #134: loss3/loss10 = 0.00159171 (* 0.0909091 = 0.000144701 loss) | |
| I0510 16:12:59.801323 10926 solver.cpp:245] Train net output #135: loss3/loss11 = 0.00127474 (* 0.0909091 = 0.000115886 loss) | |
| I0510 16:12:59.801338 10926 solver.cpp:245] Train net output #136: loss3/loss12 = 0.000898094 (* 0.0909091 = 8.16449e-05 loss) | |
| I0510 16:12:59.801352 10926 solver.cpp:245] Train net output #137: loss3/loss13 = 0.000608985 (* 0.0909091 = 5.53623e-05 loss) | |
| I0510 16:12:59.801367 10926 solver.cpp:245] Train net output #138: loss3/loss14 = 0.000587254 (* 0.0909091 = 5.33868e-05 loss) | |
| I0510 16:12:59.801380 10926 solver.cpp:245] Train net output #139: loss3/loss15 = 0.00054432 (* 0.0909091 = 4.94836e-05 loss) | |
| I0510 16:12:59.801394 10926 solver.cpp:245] Train net output #140: loss3/loss16 = 0.000418709 (* 0.0909091 = 3.80645e-05 loss) | |
| I0510 16:12:59.801409 10926 solver.cpp:245] Train net output #141: loss3/loss17 = 0.000395522 (* 0.0909091 = 3.59565e-05 loss) | |
| I0510 16:12:59.801424 10926 solver.cpp:245] Train net output #142: loss3/loss18 = 0.000573173 (* 0.0909091 = 5.21067e-05 loss) | |
| I0510 16:12:59.801437 10926 solver.cpp:245] Train net output #143: loss3/loss19 = 0.000272979 (* 0.0909091 = 2.48162e-05 loss) | |
| I0510 16:12:59.801451 10926 solver.cpp:245] Train net output #144: loss3/loss20 = 0.000525399 (* 0.0909091 = 4.77635e-05 loss) | |
| I0510 16:12:59.801465 10926 solver.cpp:245] Train net output #145: loss3/loss21 = 0.000407652 (* 0.0909091 = 3.70593e-05 loss) | |
| I0510 16:12:59.801479 10926 solver.cpp:245] Train net output #146: loss3/loss22 = 0.000238762 (* 0.0909091 = 2.17056e-05 loss) | |
| I0510 16:12:59.801503 10926 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0510 16:12:59.801517 10926 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0510 16:12:59.801527 10926 solver.cpp:245] Train net output #149: total_confidence = 2.03104e-05 | |
| I0510 16:12:59.801540 10926 solver.cpp:245] Train net output #150: total_confidence_not_rec = 6.79114e-05 | |
| I0510 16:12:59.801553 10926 sgd_solver.cpp:106] Iteration 9500, lr = 0.001 | |
| I0510 16:13:31.555196 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 32.6536 > 30) by scale factor 0.918733 | |
| I0510 16:14:20.236389 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 30.9386 > 30) by scale factor 0.969662 | |
| I0510 16:15:15.123539 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 30.2997 > 30) by scale factor 0.990109 | |
| I0510 16:15:15.417742 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 33.6137 > 30) by scale factor 0.892494 | |
| I0510 16:15:26.266685 10926 solver.cpp:456] Snapshotting to binary proto file /mnt/snapshots/mixed_lstm15_bn_iter_10000.caffemodel | |
| I0510 16:15:26.690098 10926 sgd_solver.cpp:273] Snapshotting solver state to binary proto file /mnt/snapshots/mixed_lstm15_bn_iter_10000.solverstate | |
| I0510 16:15:26.890548 10926 solver.cpp:338] Iteration 10000, Testing net (#0) | |
| I0510 16:16:09.850675 10926 solver.cpp:393] Test loss: 9.89892 | |
| I0510 16:16:09.850801 10926 solver.cpp:406] Test net output #0: loss1/accuracy = 0.0572453 | |
| I0510 16:16:09.850821 10926 solver.cpp:406] Test net output #1: loss1/accuracy01 = 0.105 | |
| I0510 16:16:09.850836 10926 solver.cpp:406] Test net output #2: loss1/accuracy02 = 0.088 | |
| I0510 16:16:09.850848 10926 solver.cpp:406] Test net output #3: loss1/accuracy03 = 0.064 | |
| I0510 16:16:09.850862 10926 solver.cpp:406] Test net output #4: loss1/accuracy04 = 0.142 | |
| I0510 16:16:09.850877 10926 solver.cpp:406] Test net output #5: loss1/accuracy05 = 0.292 | |
| I0510 16:16:09.850890 10926 solver.cpp:406] Test net output #6: loss1/accuracy06 = 0.45 | |
| I0510 16:16:09.850903 10926 solver.cpp:406] Test net output #7: loss1/accuracy07 = 0.721 | |
| I0510 16:16:09.850916 10926 solver.cpp:406] Test net output #8: loss1/accuracy08 = 0.909 | |
| I0510 16:16:09.850929 10926 solver.cpp:406] Test net output #9: loss1/accuracy09 = 0.989 | |
| I0510 16:16:09.850941 10926 solver.cpp:406] Test net output #10: loss1/accuracy10 = 0.999 | |
| I0510 16:16:09.850955 10926 solver.cpp:406] Test net output #11: loss1/accuracy11 = 1 | |
| I0510 16:16:09.850967 10926 solver.cpp:406] Test net output #12: loss1/accuracy12 = 1 | |
| I0510 16:16:09.850980 10926 solver.cpp:406] Test net output #13: loss1/accuracy13 = 1 | |
| I0510 16:16:09.850991 10926 solver.cpp:406] Test net output #14: loss1/accuracy14 = 1 | |
| I0510 16:16:09.851002 10926 solver.cpp:406] Test net output #15: loss1/accuracy15 = 1 | |
| I0510 16:16:09.851014 10926 solver.cpp:406] Test net output #16: loss1/accuracy16 = 1 | |
| I0510 16:16:09.851025 10926 solver.cpp:406] Test net output #17: loss1/accuracy17 = 1 | |
| I0510 16:16:09.851037 10926 solver.cpp:406] Test net output #18: loss1/accuracy18 = 1 | |
| I0510 16:16:09.851048 10926 solver.cpp:406] Test net output #19: loss1/accuracy19 = 1 | |
| I0510 16:16:09.851059 10926 solver.cpp:406] Test net output #20: loss1/accuracy20 = 1 | |
| I0510 16:16:09.851071 10926 solver.cpp:406] Test net output #21: loss1/accuracy21 = 1 | |
| I0510 16:16:09.851083 10926 solver.cpp:406] Test net output #22: loss1/accuracy22 = 1 | |
| I0510 16:16:09.851094 10926 solver.cpp:406] Test net output #23: loss1/accuracy_incl_empty = 0.761319 | |
| I0510 16:16:09.851106 10926 solver.cpp:406] Test net output #24: loss1/accuracy_top3 = 0.200757 | |
| I0510 16:16:09.851122 10926 solver.cpp:406] Test net output #25: loss1/cross_entropy_loss = 3.85221 (* 0.3 = 1.15566 loss) | |
| I0510 16:16:09.851137 10926 solver.cpp:406] Test net output #26: loss1/cross_entropy_loss_incl_empty = 1.0076 (* 0.3 = 0.302279 loss) | |
| I0510 16:16:09.851151 10926 solver.cpp:406] Test net output #27: loss1/loss01 = 3.31594 (* 0.0272727 = 0.0904349 loss) | |
| I0510 16:16:09.851166 10926 solver.cpp:406] Test net output #28: loss1/loss02 = 3.43003 (* 0.0272727 = 0.0935463 loss) | |
| I0510 16:16:09.851179 10926 solver.cpp:406] Test net output #29: loss1/loss03 = 3.51208 (* 0.0272727 = 0.0957839 loss) | |
| I0510 16:16:09.851193 10926 solver.cpp:406] Test net output #30: loss1/loss04 = 3.33344 (* 0.0272727 = 0.0909121 loss) | |
| I0510 16:16:09.851207 10926 solver.cpp:406] Test net output #31: loss1/loss05 = 2.88933 (* 0.0272727 = 0.0788 loss) | |
| I0510 16:16:09.851222 10926 solver.cpp:406] Test net output #32: loss1/loss06 = 2.45399 (* 0.0272727 = 0.0669269 loss) | |
| I0510 16:16:09.851234 10926 solver.cpp:406] Test net output #33: loss1/loss07 = 1.52688 (* 0.0272727 = 0.0416421 loss) | |
| I0510 16:16:09.851248 10926 solver.cpp:406] Test net output #34: loss1/loss08 = 0.562033 (* 0.0272727 = 0.0153282 loss) | |
| I0510 16:16:09.851263 10926 solver.cpp:406] Test net output #35: loss1/loss09 = 0.0935474 (* 0.0272727 = 0.00255129 loss) | |
| I0510 16:16:09.851277 10926 solver.cpp:406] Test net output #36: loss1/loss10 = 0.0282647 (* 0.0272727 = 0.000770856 loss) | |
| I0510 16:16:09.851291 10926 solver.cpp:406] Test net output #37: loss1/loss11 = 0.0189143 (* 0.0272727 = 0.000515844 loss) | |
| I0510 16:16:09.851305 10926 solver.cpp:406] Test net output #38: loss1/loss12 = 0.0168255 (* 0.0272727 = 0.000458876 loss) | |
| I0510 16:16:09.851318 10926 solver.cpp:406] Test net output #39: loss1/loss13 = 0.0130153 (* 0.0272727 = 0.000354962 loss) | |
| I0510 16:16:09.851352 10926 solver.cpp:406] Test net output #40: loss1/loss14 = 0.0113179 (* 0.0272727 = 0.00030867 loss) | |
| I0510 16:16:09.851368 10926 solver.cpp:406] Test net output #41: loss1/loss15 = 0.0106516 (* 0.0272727 = 0.000290498 loss) | |
| I0510 16:16:09.851382 10926 solver.cpp:406] Test net output #42: loss1/loss16 = 0.00956532 (* 0.0272727 = 0.000260872 loss) | |
| I0510 16:16:09.851397 10926 solver.cpp:406] Test net output #43: loss1/loss17 = 0.0075883 (* 0.0272727 = 0.000206954 loss) | |
| I0510 16:16:09.851410 10926 solver.cpp:406] Test net output #44: loss1/loss18 = 0.0063544 (* 0.0272727 = 0.000173302 loss) | |
| I0510 16:16:09.851424 10926 solver.cpp:406] Test net output #45: loss1/loss19 = 0.00459499 (* 0.0272727 = 0.000125318 loss) | |
| I0510 16:16:09.851438 10926 solver.cpp:406] Test net output #46: loss1/loss20 = 0.00513636 (* 0.0272727 = 0.000140083 loss) | |
| I0510 16:16:09.851452 10926 solver.cpp:406] Test net output #47: loss1/loss21 = 0.00458722 (* 0.0272727 = 0.000125106 loss) | |
| I0510 16:16:09.851466 10926 solver.cpp:406] Test net output #48: loss1/loss22 = 0.00433246 (* 0.0272727 = 0.000118158 loss) | |
| I0510 16:16:09.851478 10926 solver.cpp:406] Test net output #49: loss2/accuracy = 0.0632766 | |
| I0510 16:16:09.851491 10926 solver.cpp:406] Test net output #50: loss2/accuracy01 = 0.104 | |
| I0510 16:16:09.851502 10926 solver.cpp:406] Test net output #51: loss2/accuracy02 = 0.086 | |
| I0510 16:16:09.851514 10926 solver.cpp:406] Test net output #52: loss2/accuracy03 = 0.058 | |
| I0510 16:16:09.851526 10926 solver.cpp:406] Test net output #53: loss2/accuracy04 = 0.151 | |
| I0510 16:16:09.851538 10926 solver.cpp:406] Test net output #54: loss2/accuracy05 = 0.285 | |
| I0510 16:16:09.851549 10926 solver.cpp:406] Test net output #55: loss2/accuracy06 = 0.45 | |
| I0510 16:16:09.851562 10926 solver.cpp:406] Test net output #56: loss2/accuracy07 = 0.721 | |
| I0510 16:16:09.851573 10926 solver.cpp:406] Test net output #57: loss2/accuracy08 = 0.909 | |
| I0510 16:16:09.851585 10926 solver.cpp:406] Test net output #58: loss2/accuracy09 = 0.989 | |
| I0510 16:16:09.851596 10926 solver.cpp:406] Test net output #59: loss2/accuracy10 = 0.999 | |
| I0510 16:16:09.851609 10926 solver.cpp:406] Test net output #60: loss2/accuracy11 = 1 | |
| I0510 16:16:09.851620 10926 solver.cpp:406] Test net output #61: loss2/accuracy12 = 1 | |
| I0510 16:16:09.851632 10926 solver.cpp:406] Test net output #62: loss2/accuracy13 = 1 | |
| I0510 16:16:09.851644 10926 solver.cpp:406] Test net output #63: loss2/accuracy14 = 1 | |
| I0510 16:16:09.851655 10926 solver.cpp:406] Test net output #64: loss2/accuracy15 = 1 | |
| I0510 16:16:09.851666 10926 solver.cpp:406] Test net output #65: loss2/accuracy16 = 1 | |
| I0510 16:16:09.851678 10926 solver.cpp:406] Test net output #66: loss2/accuracy17 = 1 | |
| I0510 16:16:09.851689 10926 solver.cpp:406] Test net output #67: loss2/accuracy18 = 1 | |
| I0510 16:16:09.851701 10926 solver.cpp:406] Test net output #68: loss2/accuracy19 = 1 | |
| I0510 16:16:09.851712 10926 solver.cpp:406] Test net output #69: loss2/accuracy20 = 1 | |
| I0510 16:16:09.851723 10926 solver.cpp:406] Test net output #70: loss2/accuracy21 = 1 | |
| I0510 16:16:09.851735 10926 solver.cpp:406] Test net output #71: loss2/accuracy22 = 1 | |
| I0510 16:16:09.851747 10926 solver.cpp:406] Test net output #72: loss2/accuracy_incl_empty = 0.762455 | |
| I0510 16:16:09.851758 10926 solver.cpp:406] Test net output #73: loss2/accuracy_top3 = 0.211471 | |
| I0510 16:16:09.851773 10926 solver.cpp:406] Test net output #74: loss2/cross_entropy_loss = 3.73137 (* 0.3 = 1.11941 loss) | |
| I0510 16:16:09.851786 10926 solver.cpp:406] Test net output #75: loss2/cross_entropy_loss_incl_empty = 0.979519 (* 0.3 = 0.293856 loss) | |
| I0510 16:16:09.851800 10926 solver.cpp:406] Test net output #76: loss2/loss01 = 3.26388 (* 0.0272727 = 0.089015 loss) | |
| I0510 16:16:09.851817 10926 solver.cpp:406] Test net output #77: loss2/loss02 = 3.34861 (* 0.0272727 = 0.0913258 loss) | |
| I0510 16:16:09.851831 10926 solver.cpp:406] Test net output #78: loss2/loss03 = 3.45053 (* 0.0272727 = 0.0941054 loss) | |
| I0510 16:16:09.851857 10926 solver.cpp:406] Test net output #79: loss2/loss04 = 3.27438 (* 0.0272727 = 0.0893011 loss) | |
| I0510 16:16:09.851872 10926 solver.cpp:406] Test net output #80: loss2/loss05 = 2.84809 (* 0.0272727 = 0.0776753 loss) | |
| I0510 16:16:09.851886 10926 solver.cpp:406] Test net output #81: loss2/loss06 = 2.40528 (* 0.0272727 = 0.0655987 loss) | |
| I0510 16:16:09.851900 10926 solver.cpp:406] Test net output #82: loss2/loss07 = 1.5092 (* 0.0272727 = 0.0411601 loss) | |
| I0510 16:16:09.851913 10926 solver.cpp:406] Test net output #83: loss2/loss08 = 0.57438 (* 0.0272727 = 0.0156649 loss) | |
| I0510 16:16:09.851930 10926 solver.cpp:406] Test net output #84: loss2/loss09 = 0.0888605 (* 0.0272727 = 0.00242347 loss) | |
| I0510 16:16:09.851945 10926 solver.cpp:406] Test net output #85: loss2/loss10 = 0.0215383 (* 0.0272727 = 0.000587407 loss) | |
| I0510 16:16:09.851959 10926 solver.cpp:406] Test net output #86: loss2/loss11 = 0.0140181 (* 0.0272727 = 0.000382313 loss) | |
| I0510 16:16:09.851974 10926 solver.cpp:406] Test net output #87: loss2/loss12 = 0.0126228 (* 0.0272727 = 0.000344259 loss) | |
| I0510 16:16:09.851986 10926 solver.cpp:406] Test net output #88: loss2/loss13 = 0.010729 (* 0.0272727 = 0.00029261 loss) | |
| I0510 16:16:09.852000 10926 solver.cpp:406] Test net output #89: loss2/loss14 = 0.0101452 (* 0.0272727 = 0.000276688 loss) | |
| I0510 16:16:09.852015 10926 solver.cpp:406] Test net output #90: loss2/loss15 = 0.00860279 (* 0.0272727 = 0.000234622 loss) | |
| I0510 16:16:09.852028 10926 solver.cpp:406] Test net output #91: loss2/loss16 = 0.00652532 (* 0.0272727 = 0.000177963 loss) | |
| I0510 16:16:09.852042 10926 solver.cpp:406] Test net output #92: loss2/loss17 = 0.00436486 (* 0.0272727 = 0.000119042 loss) | |
| I0510 16:16:09.852056 10926 solver.cpp:406] Test net output #93: loss2/loss18 = 0.00407525 (* 0.0272727 = 0.000111143 loss) | |
| I0510 16:16:09.852069 10926 solver.cpp:406] Test net output #94: loss2/loss19 = 0.00462596 (* 0.0272727 = 0.000126162 loss) | |
| I0510 16:16:09.852083 10926 solver.cpp:406] Test net output #95: loss2/loss20 = 0.00404994 (* 0.0272727 = 0.000110453 loss) | |
| I0510 16:16:09.852097 10926 solver.cpp:406] Test net output #96: loss2/loss21 = 0.00355552 (* 0.0272727 = 9.69686e-05 loss) | |
| I0510 16:16:09.852110 10926 solver.cpp:406] Test net output #97: loss2/loss22 = 0.00413631 (* 0.0272727 = 0.000112808 loss) | |
| I0510 16:16:09.852124 10926 solver.cpp:406] Test net output #98: loss3/accuracy = 0.0706794 | |
| I0510 16:16:09.852136 10926 solver.cpp:406] Test net output #99: loss3/accuracy01 = 0.125 | |
| I0510 16:16:09.852149 10926 solver.cpp:406] Test net output #100: loss3/accuracy02 = 0.079 | |
| I0510 16:16:09.852160 10926 solver.cpp:406] Test net output #101: loss3/accuracy03 = 0.085 | |
| I0510 16:16:09.852171 10926 solver.cpp:406] Test net output #102: loss3/accuracy04 = 0.164 | |
| I0510 16:16:09.852183 10926 solver.cpp:406] Test net output #103: loss3/accuracy05 = 0.31 | |
| I0510 16:16:09.852195 10926 solver.cpp:406] Test net output #104: loss3/accuracy06 = 0.448 | |
| I0510 16:16:09.852206 10926 solver.cpp:406] Test net output #105: loss3/accuracy07 = 0.718 | |
| I0510 16:16:09.852218 10926 solver.cpp:406] Test net output #106: loss3/accuracy08 = 0.908 | |
| I0510 16:16:09.852229 10926 solver.cpp:406] Test net output #107: loss3/accuracy09 = 0.988 | |
| I0510 16:16:09.852241 10926 solver.cpp:406] Test net output #108: loss3/accuracy10 = 0.999 | |
| I0510 16:16:09.852253 10926 solver.cpp:406] Test net output #109: loss3/accuracy11 = 1 | |
| I0510 16:16:09.852264 10926 solver.cpp:406] Test net output #110: loss3/accuracy12 = 1 | |
| I0510 16:16:09.852275 10926 solver.cpp:406] Test net output #111: loss3/accuracy13 = 1 | |
| I0510 16:16:09.852288 10926 solver.cpp:406] Test net output #112: loss3/accuracy14 = 1 | |
| I0510 16:16:09.852298 10926 solver.cpp:406] Test net output #113: loss3/accuracy15 = 1 | |
| I0510 16:16:09.852309 10926 solver.cpp:406] Test net output #114: loss3/accuracy16 = 1 | |
| I0510 16:16:09.852320 10926 solver.cpp:406] Test net output #115: loss3/accuracy17 = 1 | |
| I0510 16:16:09.852342 10926 solver.cpp:406] Test net output #116: loss3/accuracy18 = 1 | |
| I0510 16:16:09.852355 10926 solver.cpp:406] Test net output #117: loss3/accuracy19 = 1 | |
| I0510 16:16:09.852367 10926 solver.cpp:406] Test net output #118: loss3/accuracy20 = 1 | |
| I0510 16:16:09.852378 10926 solver.cpp:406] Test net output #119: loss3/accuracy21 = 1 | |
| I0510 16:16:09.852391 10926 solver.cpp:406] Test net output #120: loss3/accuracy22 = 1 | |
| I0510 16:16:09.852401 10926 solver.cpp:406] Test net output #121: loss3/accuracy_incl_empty = 0.763091 | |
| I0510 16:16:09.852413 10926 solver.cpp:406] Test net output #122: loss3/accuracy_top3 = 0.236376 | |
| I0510 16:16:09.852427 10926 solver.cpp:406] Test net output #123: loss3/cross_entropy_loss = 3.27246 (* 1 = 3.27246 loss) | |
| I0510 16:16:09.852442 10926 solver.cpp:406] Test net output #124: loss3/cross_entropy_loss_incl_empty = 0.870674 (* 1 = 0.870674 loss) | |
| I0510 16:16:09.852455 10926 solver.cpp:406] Test net output #125: loss3/loss01 = 2.98682 (* 0.0909091 = 0.271529 loss) | |
| I0510 16:16:09.852469 10926 solver.cpp:406] Test net output #126: loss3/loss02 = 3.14586 (* 0.0909091 = 0.285988 loss) | |
| I0510 16:16:09.852483 10926 solver.cpp:406] Test net output #127: loss3/loss03 = 3.21011 (* 0.0909091 = 0.291828 loss) | |
| I0510 16:16:09.852496 10926 solver.cpp:406] Test net output #128: loss3/loss04 = 3.01004 (* 0.0909091 = 0.27364 loss) | |
| I0510 16:16:09.852509 10926 solver.cpp:406] Test net output #129: loss3/loss05 = 2.55737 (* 0.0909091 = 0.232488 loss) | |
| I0510 16:16:09.852524 10926 solver.cpp:406] Test net output #130: loss3/loss06 = 2.16872 (* 0.0909091 = 0.197157 loss) | |
| I0510 16:16:09.852536 10926 solver.cpp:406] Test net output #131: loss3/loss07 = 1.3465 (* 0.0909091 = 0.122409 loss) | |
| I0510 16:16:09.852550 10926 solver.cpp:406] Test net output #132: loss3/loss08 = 0.512013 (* 0.0909091 = 0.0465467 loss) | |
| I0510 16:16:09.852565 10926 solver.cpp:406] Test net output #133: loss3/loss09 = 0.0872527 (* 0.0909091 = 0.00793206 loss) | |
| I0510 16:16:09.852578 10926 solver.cpp:406] Test net output #134: loss3/loss10 = 0.0188482 (* 0.0909091 = 0.00171347 loss) | |
| I0510 16:16:09.852592 10926 solver.cpp:406] Test net output #135: loss3/loss11 = 0.0118049 (* 0.0909091 = 0.00107318 loss) | |
| I0510 16:16:09.852607 10926 solver.cpp:406] Test net output #136: loss3/loss12 = 0.00841505 (* 0.0909091 = 0.000765004 loss) | |
| I0510 16:16:09.852619 10926 solver.cpp:406] Test net output #137: loss3/loss13 = 0.00674187 (* 0.0909091 = 0.000612897 loss) | |
| I0510 16:16:09.852632 10926 solver.cpp:406] Test net output #138: loss3/loss14 = 0.00540735 (* 0.0909091 = 0.000491577 loss) | |
| I0510 16:16:09.852646 10926 solver.cpp:406] Test net output #139: loss3/loss15 = 0.00406184 (* 0.0909091 = 0.000369258 loss) | |
| I0510 16:16:09.852660 10926 solver.cpp:406] Test net output #140: loss3/loss16 = 0.0031603 (* 0.0909091 = 0.0002873 loss) | |
| I0510 16:16:09.852675 10926 solver.cpp:406] Test net output #141: loss3/loss17 = 0.00203996 (* 0.0909091 = 0.000185451 loss) | |
| I0510 16:16:09.852689 10926 solver.cpp:406] Test net output #142: loss3/loss18 = 0.00160847 (* 0.0909091 = 0.000146225 loss) | |
| I0510 16:16:09.852702 10926 solver.cpp:406] Test net output #143: loss3/loss19 = 0.00149596 (* 0.0909091 = 0.000135997 loss) | |
| I0510 16:16:09.852716 10926 solver.cpp:406] Test net output #144: loss3/loss20 = 0.00103426 (* 0.0909091 = 9.40236e-05 loss) | |
| I0510 16:16:09.852731 10926 solver.cpp:406] Test net output #145: loss3/loss21 = 0.000943762 (* 0.0909091 = 8.57966e-05 loss) | |
| I0510 16:16:09.852746 10926 solver.cpp:406] Test net output #146: loss3/loss22 = 0.000901769 (* 0.0909091 = 8.1979e-05 loss) | |
| I0510 16:16:09.852756 10926 solver.cpp:406] Test net output #147: total_accuracy = 0 | |
| I0510 16:16:09.852768 10926 solver.cpp:406] Test net output #148: total_accuracy_not_rec = 0 | |
| I0510 16:16:09.852779 10926 solver.cpp:406] Test net output #149: total_confidence = 0.000175639 | |
| I0510 16:16:09.852787 10926 solver.cpp:406] Test net output #150: total_confidence_not_rec = 0.000170115 | |
| I0510 16:16:09.852807 10926 solver.cpp:338] Iteration 10000, Testing net (#1) | |
| I0510 16:16:52.709969 10926 solver.cpp:393] Test loss: 10.6944 | |
| I0510 16:16:52.710098 10926 solver.cpp:406] Test net output #0: loss1/accuracy = 0.0596216 | |
| I0510 16:16:52.710119 10926 solver.cpp:406] Test net output #1: loss1/accuracy01 = 0.11 | |
| I0510 16:16:52.710134 10926 solver.cpp:406] Test net output #2: loss1/accuracy02 = 0.087 | |
| I0510 16:16:52.710147 10926 solver.cpp:406] Test net output #3: loss1/accuracy03 = 0.061 | |
| I0510 16:16:52.710160 10926 solver.cpp:406] Test net output #4: loss1/accuracy04 = 0.121 | |
| I0510 16:16:52.710173 10926 solver.cpp:406] Test net output #5: loss1/accuracy05 = 0.278 | |
| I0510 16:16:52.710186 10926 solver.cpp:406] Test net output #6: loss1/accuracy06 = 0.404 | |
| I0510 16:16:52.710198 10926 solver.cpp:406] Test net output #7: loss1/accuracy07 = 0.658 | |
| I0510 16:16:52.710211 10926 solver.cpp:406] Test net output #8: loss1/accuracy08 = 0.81 | |
| I0510 16:16:52.710223 10926 solver.cpp:406] Test net output #9: loss1/accuracy09 = 0.894 | |
| I0510 16:16:52.710237 10926 solver.cpp:406] Test net output #10: loss1/accuracy10 = 0.917 | |
| I0510 16:16:52.710249 10926 solver.cpp:406] Test net output #11: loss1/accuracy11 = 0.934 | |
| I0510 16:16:52.710263 10926 solver.cpp:406] Test net output #12: loss1/accuracy12 = 0.942 | |
| I0510 16:16:52.710274 10926 solver.cpp:406] Test net output #13: loss1/accuracy13 = 0.958 | |
| I0510 16:16:52.710286 10926 solver.cpp:406] Test net output #14: loss1/accuracy14 = 0.969 | |
| I0510 16:16:52.710299 10926 solver.cpp:406] Test net output #15: loss1/accuracy15 = 0.976 | |
| I0510 16:16:52.710310 10926 solver.cpp:406] Test net output #16: loss1/accuracy16 = 0.983 | |
| I0510 16:16:52.710322 10926 solver.cpp:406] Test net output #17: loss1/accuracy17 = 0.994 | |
| I0510 16:16:52.710335 10926 solver.cpp:406] Test net output #18: loss1/accuracy18 = 0.996 | |
| I0510 16:16:52.710346 10926 solver.cpp:406] Test net output #19: loss1/accuracy19 = 0.997 | |
| I0510 16:16:52.710358 10926 solver.cpp:406] Test net output #20: loss1/accuracy20 = 0.997 | |
| I0510 16:16:52.710371 10926 solver.cpp:406] Test net output #21: loss1/accuracy21 = 0.998 | |
| I0510 16:16:52.710382 10926 solver.cpp:406] Test net output #22: loss1/accuracy22 = 0.999 | |
| I0510 16:16:52.710394 10926 solver.cpp:406] Test net output #23: loss1/accuracy_incl_empty = 0.731364 | |
| I0510 16:16:52.710407 10926 solver.cpp:406] Test net output #24: loss1/accuracy_top3 = 0.209275 | |
| I0510 16:16:52.710422 10926 solver.cpp:406] Test net output #25: loss1/cross_entropy_loss = 3.90746 (* 0.3 = 1.17224 loss) | |
| I0510 16:16:52.710438 10926 solver.cpp:406] Test net output #26: loss1/cross_entropy_loss_incl_empty = 1.16254 (* 0.3 = 0.348763 loss) | |
| I0510 16:16:52.710453 10926 solver.cpp:406] Test net output #27: loss1/loss01 = 3.32083 (* 0.0272727 = 0.0905681 loss) | |
| I0510 16:16:52.710466 10926 solver.cpp:406] Test net output #28: loss1/loss02 = 3.39654 (* 0.0272727 = 0.092633 loss) | |
| I0510 16:16:52.710480 10926 solver.cpp:406] Test net output #29: loss1/loss03 = 3.54489 (* 0.0272727 = 0.0966788 loss) | |
| I0510 16:16:52.710494 10926 solver.cpp:406] Test net output #30: loss1/loss04 = 3.43563 (* 0.0272727 = 0.093699 loss) | |
| I0510 16:16:52.710507 10926 solver.cpp:406] Test net output #31: loss1/loss05 = 3.0199 (* 0.0272727 = 0.0823608 loss) | |
| I0510 16:16:52.710521 10926 solver.cpp:406] Test net output #32: loss1/loss06 = 2.56816 (* 0.0272727 = 0.0700406 loss) | |
| I0510 16:16:52.710536 10926 solver.cpp:406] Test net output #33: loss1/loss07 = 1.69533 (* 0.0272727 = 0.0462364 loss) | |
| I0510 16:16:52.710548 10926 solver.cpp:406] Test net output #34: loss1/loss08 = 0.973701 (* 0.0272727 = 0.0265555 loss) | |
| I0510 16:16:52.710562 10926 solver.cpp:406] Test net output #35: loss1/loss09 = 0.513662 (* 0.0272727 = 0.014009 loss) | |
| I0510 16:16:52.710577 10926 solver.cpp:406] Test net output #36: loss1/loss10 = 0.441548 (* 0.0272727 = 0.0120422 loss) | |
| I0510 16:16:52.710592 10926 solver.cpp:406] Test net output #37: loss1/loss11 = 0.366404 (* 0.0272727 = 0.00999283 loss) | |
| I0510 16:16:52.710605 10926 solver.cpp:406] Test net output #38: loss1/loss12 = 0.337548 (* 0.0272727 = 0.00920586 loss) | |
| I0510 16:16:52.710639 10926 solver.cpp:406] Test net output #39: loss1/loss13 = 0.250481 (* 0.0272727 = 0.0068313 loss) | |
| I0510 16:16:52.710654 10926 solver.cpp:406] Test net output #40: loss1/loss14 = 0.202369 (* 0.0272727 = 0.00551915 loss) | |
| I0510 16:16:52.710669 10926 solver.cpp:406] Test net output #41: loss1/loss15 = 0.158796 (* 0.0272727 = 0.00433081 loss) | |
| I0510 16:16:52.710682 10926 solver.cpp:406] Test net output #42: loss1/loss16 = 0.128961 (* 0.0272727 = 0.00351712 loss) | |
| I0510 16:16:52.710697 10926 solver.cpp:406] Test net output #43: loss1/loss17 = 0.0544092 (* 0.0272727 = 0.00148389 loss) | |
| I0510 16:16:52.710711 10926 solver.cpp:406] Test net output #44: loss1/loss18 = 0.0407171 (* 0.0272727 = 0.00111047 loss) | |
| I0510 16:16:52.710726 10926 solver.cpp:406] Test net output #45: loss1/loss19 = 0.0309337 (* 0.0272727 = 0.000843646 loss) | |
| I0510 16:16:52.710739 10926 solver.cpp:406] Test net output #46: loss1/loss20 = 0.0306705 (* 0.0272727 = 0.000836469 loss) | |
| I0510 16:16:52.710753 10926 solver.cpp:406] Test net output #47: loss1/loss21 = 0.0218988 (* 0.0272727 = 0.000597239 loss) | |
| I0510 16:16:52.710767 10926 solver.cpp:406] Test net output #48: loss1/loss22 = 0.0137546 (* 0.0272727 = 0.000375125 loss) | |
| I0510 16:16:52.710779 10926 solver.cpp:406] Test net output #49: loss2/accuracy = 0.0608505 | |
| I0510 16:16:52.710793 10926 solver.cpp:406] Test net output #50: loss2/accuracy01 = 0.113 | |
| I0510 16:16:52.710804 10926 solver.cpp:406] Test net output #51: loss2/accuracy02 = 0.094 | |
| I0510 16:16:52.710816 10926 solver.cpp:406] Test net output #52: loss2/accuracy03 = 0.059 | |
| I0510 16:16:52.710827 10926 solver.cpp:406] Test net output #53: loss2/accuracy04 = 0.133 | |
| I0510 16:16:52.710839 10926 solver.cpp:406] Test net output #54: loss2/accuracy05 = 0.27 | |
| I0510 16:16:52.710851 10926 solver.cpp:406] Test net output #55: loss2/accuracy06 = 0.405 | |
| I0510 16:16:52.710863 10926 solver.cpp:406] Test net output #56: loss2/accuracy07 = 0.658 | |
| I0510 16:16:52.710878 10926 solver.cpp:406] Test net output #57: loss2/accuracy08 = 0.809 | |
| I0510 16:16:52.710891 10926 solver.cpp:406] Test net output #58: loss2/accuracy09 = 0.894 | |
| I0510 16:16:52.710903 10926 solver.cpp:406] Test net output #59: loss2/accuracy10 = 0.917 | |
| I0510 16:16:52.710916 10926 solver.cpp:406] Test net output #60: loss2/accuracy11 = 0.934 | |
| I0510 16:16:52.710927 10926 solver.cpp:406] Test net output #61: loss2/accuracy12 = 0.942 | |
| I0510 16:16:52.710938 10926 solver.cpp:406] Test net output #62: loss2/accuracy13 = 0.958 | |
| I0510 16:16:52.710950 10926 solver.cpp:406] Test net output #63: loss2/accuracy14 = 0.969 | |
| I0510 16:16:52.710961 10926 solver.cpp:406] Test net output #64: loss2/accuracy15 = 0.976 | |
| I0510 16:16:52.710973 10926 solver.cpp:406] Test net output #65: loss2/accuracy16 = 0.983 | |
| I0510 16:16:52.710985 10926 solver.cpp:406] Test net output #66: loss2/accuracy17 = 0.994 | |
| I0510 16:16:52.710997 10926 solver.cpp:406] Test net output #67: loss2/accuracy18 = 0.996 | |
| I0510 16:16:52.711009 10926 solver.cpp:406] Test net output #68: loss2/accuracy19 = 0.997 | |
| I0510 16:16:52.711020 10926 solver.cpp:406] Test net output #69: loss2/accuracy20 = 0.997 | |
| I0510 16:16:52.711031 10926 solver.cpp:406] Test net output #70: loss2/accuracy21 = 0.998 | |
| I0510 16:16:52.711043 10926 solver.cpp:406] Test net output #71: loss2/accuracy22 = 0.999 | |
| I0510 16:16:52.711055 10926 solver.cpp:406] Test net output #72: loss2/accuracy_incl_empty = 0.731046 | |
| I0510 16:16:52.711067 10926 solver.cpp:406] Test net output #73: loss2/accuracy_top3 = 0.219391 | |
| I0510 16:16:52.711081 10926 solver.cpp:406] Test net output #74: loss2/cross_entropy_loss = 3.8005 (* 0.3 = 1.14015 loss) | |
| I0510 16:16:52.711096 10926 solver.cpp:406] Test net output #75: loss2/cross_entropy_loss_incl_empty = 1.13556 (* 0.3 = 0.340667 loss) | |
| I0510 16:16:52.711108 10926 solver.cpp:406] Test net output #76: loss2/loss01 = 3.25543 (* 0.0272727 = 0.0887846 loss) | |
| I0510 16:16:52.711125 10926 solver.cpp:406] Test net output #77: loss2/loss02 = 3.33644 (* 0.0272727 = 0.0909937 loss) | |
| I0510 16:16:52.711151 10926 solver.cpp:406] Test net output #78: loss2/loss03 = 3.49792 (* 0.0272727 = 0.0953977 loss) | |
| I0510 16:16:52.711168 10926 solver.cpp:406] Test net output #79: loss2/loss04 = 3.38095 (* 0.0272727 = 0.0922078 loss) | |
| I0510 16:16:52.711180 10926 solver.cpp:406] Test net output #80: loss2/loss05 = 2.98997 (* 0.0272727 = 0.0815446 loss) | |
| I0510 16:16:52.711194 10926 solver.cpp:406] Test net output #81: loss2/loss06 = 2.53271 (* 0.0272727 = 0.0690739 loss) | |
| I0510 16:16:52.711208 10926 solver.cpp:406] Test net output #82: loss2/loss07 = 1.68412 (* 0.0272727 = 0.0459307 loss) | |
| I0510 16:16:52.711222 10926 solver.cpp:406] Test net output #83: loss2/loss08 = 0.98077 (* 0.0272727 = 0.0267483 loss) | |
| I0510 16:16:52.711236 10926 solver.cpp:406] Test net output #84: loss2/loss09 = 0.511178 (* 0.0272727 = 0.0139412 loss) | |
| I0510 16:16:52.711249 10926 solver.cpp:406] Test net output #85: loss2/loss10 = 0.423496 (* 0.0272727 = 0.0115499 loss) | |
| I0510 16:16:52.711264 10926 solver.cpp:406] Test net output #86: loss2/loss11 = 0.351096 (* 0.0272727 = 0.00957534 loss) | |
| I0510 16:16:52.711278 10926 solver.cpp:406] Test net output #87: loss2/loss12 = 0.329688 (* 0.0272727 = 0.0089915 loss) | |
| I0510 16:16:52.711292 10926 solver.cpp:406] Test net output #88: loss2/loss13 = 0.244019 (* 0.0272727 = 0.00665508 loss) | |
| I0510 16:16:52.711307 10926 solver.cpp:406] Test net output #89: loss2/loss14 = 0.190155 (* 0.0272727 = 0.00518604 loss) | |
| I0510 16:16:52.711320 10926 solver.cpp:406] Test net output #90: loss2/loss15 = 0.156163 (* 0.0272727 = 0.00425899 loss) | |
| I0510 16:16:52.711334 10926 solver.cpp:406] Test net output #91: loss2/loss16 = 0.126951 (* 0.0272727 = 0.0034623 loss) | |
| I0510 16:16:52.711349 10926 solver.cpp:406] Test net output #92: loss2/loss17 = 0.0510019 (* 0.0272727 = 0.00139096 loss) | |
| I0510 16:16:52.711362 10926 solver.cpp:406] Test net output #93: loss2/loss18 = 0.0379252 (* 0.0272727 = 0.00103432 loss) | |
| I0510 16:16:52.711376 10926 solver.cpp:406] Test net output #94: loss2/loss19 = 0.032135 (* 0.0272727 = 0.000876408 loss) | |
| I0510 16:16:52.711390 10926 solver.cpp:406] Test net output #95: loss2/loss20 = 0.0313321 (* 0.0272727 = 0.000854511 loss) | |
| I0510 16:16:52.711405 10926 solver.cpp:406] Test net output #96: loss2/loss21 = 0.0224552 (* 0.0272727 = 0.000612416 loss) | |
| I0510 16:16:52.711418 10926 solver.cpp:406] Test net output #97: loss2/loss22 = 0.0139137 (* 0.0272727 = 0.000379465 loss) | |
| I0510 16:16:52.711431 10926 solver.cpp:406] Test net output #98: loss3/accuracy = 0.0751488 | |
| I0510 16:16:52.711442 10926 solver.cpp:406] Test net output #99: loss3/accuracy01 = 0.114 | |
| I0510 16:16:52.711454 10926 solver.cpp:406] Test net output #100: loss3/accuracy02 = 0.091 | |
| I0510 16:16:52.711467 10926 solver.cpp:406] Test net output #101: loss3/accuracy03 = 0.08 | |
| I0510 16:16:52.711478 10926 solver.cpp:406] Test net output #102: loss3/accuracy04 = 0.142 | |
| I0510 16:16:52.711489 10926 solver.cpp:406] Test net output #103: loss3/accuracy05 = 0.282 | |
| I0510 16:16:52.711501 10926 solver.cpp:406] Test net output #104: loss3/accuracy06 = 0.414 | |
| I0510 16:16:52.711513 10926 solver.cpp:406] Test net output #105: loss3/accuracy07 = 0.662 | |
| I0510 16:16:52.711524 10926 solver.cpp:406] Test net output #106: loss3/accuracy08 = 0.81 | |
| I0510 16:16:52.711536 10926 solver.cpp:406] Test net output #107: loss3/accuracy09 = 0.898 | |
| I0510 16:16:52.711549 10926 solver.cpp:406] Test net output #108: loss3/accuracy10 = 0.916 | |
| I0510 16:16:52.711560 10926 solver.cpp:406] Test net output #109: loss3/accuracy11 = 0.934 | |
| I0510 16:16:52.711572 10926 solver.cpp:406] Test net output #110: loss3/accuracy12 = 0.942 | |
| I0510 16:16:52.711585 10926 solver.cpp:406] Test net output #111: loss3/accuracy13 = 0.958 | |
| I0510 16:16:52.711601 10926 solver.cpp:406] Test net output #112: loss3/accuracy14 = 0.969 | |
| I0510 16:16:52.711612 10926 solver.cpp:406] Test net output #113: loss3/accuracy15 = 0.976 | |
| I0510 16:16:52.711624 10926 solver.cpp:406] Test net output #114: loss3/accuracy16 = 0.983 | |
| I0510 16:16:52.711647 10926 solver.cpp:406] Test net output #115: loss3/accuracy17 = 0.994 | |
| I0510 16:16:52.711660 10926 solver.cpp:406] Test net output #116: loss3/accuracy18 = 0.996 | |
| I0510 16:16:52.711673 10926 solver.cpp:406] Test net output #117: loss3/accuracy19 = 0.997 | |
| I0510 16:16:52.711685 10926 solver.cpp:406] Test net output #118: loss3/accuracy20 = 0.997 | |
| I0510 16:16:52.711697 10926 solver.cpp:406] Test net output #119: loss3/accuracy21 = 0.998 | |
| I0510 16:16:52.711709 10926 solver.cpp:406] Test net output #120: loss3/accuracy22 = 0.999 | |
| I0510 16:16:52.711720 10926 solver.cpp:406] Test net output #121: loss3/accuracy_incl_empty = 0.733091 | |
| I0510 16:16:52.711732 10926 solver.cpp:406] Test net output #122: loss3/accuracy_top3 = 0.232941 | |
| I0510 16:16:52.711746 10926 solver.cpp:406] Test net output #123: loss3/cross_entropy_loss = 3.33468 (* 1 = 3.33468 loss) | |
| I0510 16:16:52.711760 10926 solver.cpp:406] Test net output #124: loss3/cross_entropy_loss_incl_empty = 1.01491 (* 1 = 1.01491 loss) | |
| I0510 16:16:52.711774 10926 solver.cpp:406] Test net output #125: loss3/loss01 = 2.9806 (* 0.0909091 = 0.270963 loss) | |
| I0510 16:16:52.711788 10926 solver.cpp:406] Test net output #126: loss3/loss02 = 3.12976 (* 0.0909091 = 0.284524 loss) | |
| I0510 16:16:52.711802 10926 solver.cpp:406] Test net output #127: loss3/loss03 = 3.23284 (* 0.0909091 = 0.293894 loss) | |
| I0510 16:16:52.711815 10926 solver.cpp:406] Test net output #128: loss3/loss04 = 3.13132 (* 0.0909091 = 0.284665 loss) | |
| I0510 16:16:52.711828 10926 solver.cpp:406] Test net output #129: loss3/loss05 = 2.71831 (* 0.0909091 = 0.247119 loss) | |
| I0510 16:16:52.711843 10926 solver.cpp:406] Test net output #130: loss3/loss06 = 2.32793 (* 0.0909091 = 0.21163 loss) | |
| I0510 16:16:52.711855 10926 solver.cpp:406] Test net output #131: loss3/loss07 = 1.50611 (* 0.0909091 = 0.136919 loss) | |
| I0510 16:16:52.711869 10926 solver.cpp:406] Test net output #132: loss3/loss08 = 0.889979 (* 0.0909091 = 0.0809072 loss) | |
| I0510 16:16:52.711882 10926 solver.cpp:406] Test net output #133: loss3/loss09 = 0.466595 (* 0.0909091 = 0.0424177 loss) | |
| I0510 16:16:52.711895 10926 solver.cpp:406] Test net output #134: loss3/loss10 = 0.392377 (* 0.0909091 = 0.0356707 loss) | |
| I0510 16:16:52.711910 10926 solver.cpp:406] Test net output #135: loss3/loss11 = 0.31422 (* 0.0909091 = 0.0285655 loss) | |
| I0510 16:16:52.711923 10926 solver.cpp:406] Test net output #136: loss3/loss12 = 0.292554 (* 0.0909091 = 0.0265958 loss) | |
| I0510 16:16:52.711941 10926 solver.cpp:406] Test net output #137: loss3/loss13 = 0.212132 (* 0.0909091 = 0.0192848 loss) | |
| I0510 16:16:52.711954 10926 solver.cpp:406] Test net output #138: loss3/loss14 = 0.159367 (* 0.0909091 = 0.0144879 loss) | |
| I0510 16:16:52.711964 10926 solver.cpp:406] Test net output #139: loss3/loss15 = 0.130622 (* 0.0909091 = 0.0118747 loss) | |
| I0510 16:16:52.711974 10926 solver.cpp:406] Test net output #140: loss3/loss16 = 0.105092 (* 0.0909091 = 0.0095538 loss) | |
| I0510 16:16:52.711983 10926 solver.cpp:406] Test net output #141: loss3/loss17 = 0.0438763 (* 0.0909091 = 0.00398875 loss) | |
| I0510 16:16:52.711998 10926 solver.cpp:406] Test net output #142: loss3/loss18 = 0.0347076 (* 0.0909091 = 0.00315524 loss) | |
| I0510 16:16:52.712013 10926 solver.cpp:406] Test net output #143: loss3/loss19 = 0.0262195 (* 0.0909091 = 0.00238359 loss) | |
| I0510 16:16:52.712026 10926 solver.cpp:406] Test net output #144: loss3/loss20 = 0.0287473 (* 0.0909091 = 0.00261339 loss) | |
| I0510 16:16:52.712041 10926 solver.cpp:406] Test net output #145: loss3/loss21 = 0.0194079 (* 0.0909091 = 0.00176435 loss) | |
| I0510 16:16:52.712054 10926 solver.cpp:406] Test net output #146: loss3/loss22 = 0.0118896 (* 0.0909091 = 0.00108087 loss) | |
| I0510 16:16:52.712066 10926 solver.cpp:406] Test net output #147: total_accuracy = 0 | |
| I0510 16:16:52.712079 10926 solver.cpp:406] Test net output #148: total_accuracy_not_rec = 0 | |
| I0510 16:16:52.712090 10926 solver.cpp:406] Test net output #149: total_confidence = 0.000158094 | |
| I0510 16:16:52.712111 10926 solver.cpp:406] Test net output #150: total_confidence_not_rec = 0.000158436 | |
| I0510 16:16:52.858057 10926 solver.cpp:229] Iteration 10000, loss = 10.6925 | |
| I0510 16:16:52.858136 10926 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0930233 | |
| I0510 16:16:52.858155 10926 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0510 16:16:52.858170 10926 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0510 16:16:52.858182 10926 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.125 | |
| I0510 16:16:52.858196 10926 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0 | |
| I0510 16:16:52.858208 10926 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.25 | |
| I0510 16:16:52.858222 10926 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.625 | |
| I0510 16:16:52.858234 10926 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.75 | |
| I0510 16:16:52.858248 10926 solver.cpp:245] Train net output #8: loss1/accuracy08 = 1 | |
| I0510 16:16:52.858270 10926 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0510 16:16:52.858297 10926 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0510 16:16:52.858319 10926 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0510 16:16:52.858333 10926 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0510 16:16:52.858345 10926 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0510 16:16:52.858358 10926 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0510 16:16:52.858371 10926 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0510 16:16:52.858382 10926 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0510 16:16:52.858394 10926 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0510 16:16:52.858407 10926 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0510 16:16:52.858419 10926 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0510 16:16:52.858431 10926 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0510 16:16:52.858443 10926 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0510 16:16:52.858455 10926 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0510 16:16:52.858467 10926 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.778409 | |
| I0510 16:16:52.858479 10926 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.232558 | |
| I0510 16:16:52.858496 10926 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.19535 (* 0.3 = 0.958604 loss) | |
| I0510 16:16:52.858511 10926 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 0.867334 (* 0.3 = 0.2602 loss) | |
| I0510 16:16:52.858526 10926 solver.cpp:245] Train net output #27: loss1/loss01 = 3.617 (* 0.0272727 = 0.0986454 loss) | |
| I0510 16:16:52.858541 10926 solver.cpp:245] Train net output #28: loss1/loss02 = 3.49081 (* 0.0272727 = 0.095204 loss) | |
| I0510 16:16:52.858556 10926 solver.cpp:245] Train net output #29: loss1/loss03 = 3.15976 (* 0.0272727 = 0.0861752 loss) | |
| I0510 16:16:52.858571 10926 solver.cpp:245] Train net output #30: loss1/loss04 = 4.15877 (* 0.0272727 = 0.113421 loss) | |
| I0510 16:16:52.858584 10926 solver.cpp:245] Train net output #31: loss1/loss05 = 3.62306 (* 0.0272727 = 0.0988107 loss) | |
| I0510 16:16:52.858598 10926 solver.cpp:245] Train net output #32: loss1/loss06 = 2.02445 (* 0.0272727 = 0.0552123 loss) | |
| I0510 16:16:52.858613 10926 solver.cpp:245] Train net output #33: loss1/loss07 = 1.27415 (* 0.0272727 = 0.0347496 loss) | |
| I0510 16:16:52.858628 10926 solver.cpp:245] Train net output #34: loss1/loss08 = 0.0854135 (* 0.0272727 = 0.00232946 loss) | |
| I0510 16:16:52.858642 10926 solver.cpp:245] Train net output #35: loss1/loss09 = 0.0174984 (* 0.0272727 = 0.000477229 loss) | |
| I0510 16:16:52.858657 10926 solver.cpp:245] Train net output #36: loss1/loss10 = 0.0145133 (* 0.0272727 = 0.000395818 loss) | |
| I0510 16:16:52.858672 10926 solver.cpp:245] Train net output #37: loss1/loss11 = 0.0193836 (* 0.0272727 = 0.000528645 loss) | |
| I0510 16:16:52.858726 10926 solver.cpp:245] Train net output #38: loss1/loss12 = 0.0192418 (* 0.0272727 = 0.000524777 loss) | |
| I0510 16:16:52.858741 10926 solver.cpp:245] Train net output #39: loss1/loss13 = 0.00742151 (* 0.0272727 = 0.000202405 loss) | |
| I0510 16:16:52.858757 10926 solver.cpp:245] Train net output #40: loss1/loss14 = 0.00775786 (* 0.0272727 = 0.000211578 loss) | |
| I0510 16:16:52.858770 10926 solver.cpp:245] Train net output #41: loss1/loss15 = 0.012796 (* 0.0272727 = 0.000348981 loss) | |
| I0510 16:16:52.858785 10926 solver.cpp:245] Train net output #42: loss1/loss16 = 0.00969915 (* 0.0272727 = 0.000264522 loss) | |
| I0510 16:16:52.858799 10926 solver.cpp:245] Train net output #43: loss1/loss17 = 0.00482746 (* 0.0272727 = 0.000131658 loss) | |
| I0510 16:16:52.858814 10926 solver.cpp:245] Train net output #44: loss1/loss18 = 0.00595108 (* 0.0272727 = 0.000162302 loss) | |
| I0510 16:16:52.858829 10926 solver.cpp:245] Train net output #45: loss1/loss19 = 0.00330016 (* 0.0272727 = 9.00044e-05 loss) | |
| I0510 16:16:52.858844 10926 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00208751 (* 0.0272727 = 5.69321e-05 loss) | |
| I0510 16:16:52.858868 10926 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00240741 (* 0.0272727 = 6.56567e-05 loss) | |
| I0510 16:16:52.858886 10926 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0026035 (* 0.0272727 = 7.10046e-05 loss) | |
| I0510 16:16:52.858899 10926 solver.cpp:245] Train net output #49: loss2/accuracy = 0 | |
| I0510 16:16:52.858911 10926 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.25 | |
| I0510 16:16:52.858924 10926 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0510 16:16:52.858937 10926 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0510 16:16:52.858948 10926 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0 | |
| I0510 16:16:52.858960 10926 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.25 | |
| I0510 16:16:52.858973 10926 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.625 | |
| I0510 16:16:52.858984 10926 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.75 | |
| I0510 16:16:52.859000 10926 solver.cpp:245] Train net output #57: loss2/accuracy08 = 1 | |
| I0510 16:16:52.859012 10926 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0510 16:16:52.859025 10926 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0510 16:16:52.859037 10926 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0510 16:16:52.859050 10926 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0510 16:16:52.859061 10926 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0510 16:16:52.859072 10926 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0510 16:16:52.859084 10926 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0510 16:16:52.859097 10926 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0510 16:16:52.859108 10926 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0510 16:16:52.859120 10926 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0510 16:16:52.859133 10926 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0510 16:16:52.859144 10926 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0510 16:16:52.859156 10926 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0510 16:16:52.859169 10926 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0510 16:16:52.859180 10926 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.75 | |
| I0510 16:16:52.859194 10926 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.255814 | |
| I0510 16:16:52.859207 10926 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.30585 (* 0.3 = 0.991754 loss) | |
| I0510 16:16:52.859221 10926 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 0.888879 (* 0.3 = 0.266664 loss) | |
| I0510 16:16:52.859236 10926 solver.cpp:245] Train net output #76: loss2/loss01 = 3.19399 (* 0.0272727 = 0.0871089 loss) | |
| I0510 16:16:52.859262 10926 solver.cpp:245] Train net output #77: loss2/loss02 = 4.32962 (* 0.0272727 = 0.118081 loss) | |
| I0510 16:16:52.859279 10926 solver.cpp:245] Train net output #78: loss2/loss03 = 3.74852 (* 0.0272727 = 0.102232 loss) | |
| I0510 16:16:52.859293 10926 solver.cpp:245] Train net output #79: loss2/loss04 = 3.81807 (* 0.0272727 = 0.104129 loss) | |
| I0510 16:16:52.859308 10926 solver.cpp:245] Train net output #80: loss2/loss05 = 3.38782 (* 0.0272727 = 0.092395 loss) | |
| I0510 16:16:52.859323 10926 solver.cpp:245] Train net output #81: loss2/loss06 = 1.73835 (* 0.0272727 = 0.0474096 loss) | |
| I0510 16:16:52.859336 10926 solver.cpp:245] Train net output #82: loss2/loss07 = 1.2896 (* 0.0272727 = 0.0351708 loss) | |
| I0510 16:16:52.859351 10926 solver.cpp:245] Train net output #83: loss2/loss08 = 0.0480425 (* 0.0272727 = 0.00131025 loss) | |
| I0510 16:16:52.859365 10926 solver.cpp:245] Train net output #84: loss2/loss09 = 0.00943947 (* 0.0272727 = 0.00025744 loss) | |
| I0510 16:16:52.859380 10926 solver.cpp:245] Train net output #85: loss2/loss10 = 0.00548136 (* 0.0272727 = 0.000149492 loss) | |
| I0510 16:16:52.859395 10926 solver.cpp:245] Train net output #86: loss2/loss11 = 0.00444309 (* 0.0272727 = 0.000121175 loss) | |
| I0510 16:16:52.859410 10926 solver.cpp:245] Train net output #87: loss2/loss12 = 0.00479322 (* 0.0272727 = 0.000130724 loss) | |
| I0510 16:16:52.859423 10926 solver.cpp:245] Train net output #88: loss2/loss13 = 0.00657957 (* 0.0272727 = 0.000179443 loss) | |
| I0510 16:16:52.859438 10926 solver.cpp:245] Train net output #89: loss2/loss14 = 0.00440979 (* 0.0272727 = 0.000120267 loss) | |
| I0510 16:16:52.859452 10926 solver.cpp:245] Train net output #90: loss2/loss15 = 0.00422748 (* 0.0272727 = 0.000115295 loss) | |
| I0510 16:16:52.859467 10926 solver.cpp:245] Train net output #91: loss2/loss16 = 0.00317757 (* 0.0272727 = 8.6661e-05 loss) | |
| I0510 16:16:52.859482 10926 solver.cpp:245] Train net output #92: loss2/loss17 = 0.00184396 (* 0.0272727 = 5.02899e-05 loss) | |
| I0510 16:16:52.859495 10926 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0015962 (* 0.0272727 = 4.35326e-05 loss) | |
| I0510 16:16:52.859510 10926 solver.cpp:245] Train net output #94: loss2/loss19 = 0.0017739 (* 0.0272727 = 4.8379e-05 loss) | |
| I0510 16:16:52.859524 10926 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00189646 (* 0.0272727 = 5.17216e-05 loss) | |
| I0510 16:16:52.859539 10926 solver.cpp:245] Train net output #96: loss2/loss21 = 0.000784563 (* 0.0272727 = 2.13972e-05 loss) | |
| I0510 16:16:52.859555 10926 solver.cpp:245] Train net output #97: loss2/loss22 = 0.000761393 (* 0.0272727 = 2.07653e-05 loss) | |
| I0510 16:16:52.859566 10926 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0697674 | |
| I0510 16:16:52.859580 10926 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.25 | |
| I0510 16:16:52.859592 10926 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.125 | |
| I0510 16:16:52.859606 10926 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0.25 | |
| I0510 16:16:52.859617 10926 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0510 16:16:52.859629 10926 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.25 | |
| I0510 16:16:52.859642 10926 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.625 | |
| I0510 16:16:52.859654 10926 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.75 | |
| I0510 16:16:52.859666 10926 solver.cpp:245] Train net output #106: loss3/accuracy08 = 1 | |
| I0510 16:16:52.859678 10926 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0510 16:16:52.859691 10926 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0510 16:16:52.859704 10926 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0510 16:16:52.859715 10926 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0510 16:16:52.859727 10926 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0510 16:16:52.859740 10926 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0510 16:16:52.859766 10926 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0510 16:16:52.859778 10926 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0510 16:16:52.859791 10926 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0510 16:16:52.859803 10926 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0510 16:16:52.859815 10926 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0510 16:16:52.859825 10926 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0510 16:16:52.859833 10926 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0510 16:16:52.859845 10926 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0510 16:16:52.859858 10926 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.767045 | |
| I0510 16:16:52.859869 10926 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.209302 | |
| I0510 16:16:52.859884 10926 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.1797 (* 1 = 3.1797 loss) | |
| I0510 16:16:52.859899 10926 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 0.841071 (* 1 = 0.841071 loss) | |
| I0510 16:16:52.859913 10926 solver.cpp:245] Train net output #125: loss3/loss01 = 3.00265 (* 0.0909091 = 0.272968 loss) | |
| I0510 16:16:52.859927 10926 solver.cpp:245] Train net output #126: loss3/loss02 = 3.31562 (* 0.0909091 = 0.30142 loss) | |
| I0510 16:16:52.859941 10926 solver.cpp:245] Train net output #127: loss3/loss03 = 2.8833 (* 0.0909091 = 0.262118 loss) | |
| I0510 16:16:52.859956 10926 solver.cpp:245] Train net output #128: loss3/loss04 = 3.32338 (* 0.0909091 = 0.302125 loss) | |
| I0510 16:16:52.859969 10926 solver.cpp:245] Train net output #129: loss3/loss05 = 2.91877 (* 0.0909091 = 0.265343 loss) | |
| I0510 16:16:52.859984 10926 solver.cpp:245] Train net output #130: loss3/loss06 = 1.51205 (* 0.0909091 = 0.137459 loss) | |
| I0510 16:16:52.859998 10926 solver.cpp:245] Train net output #131: loss3/loss07 = 1.06772 (* 0.0909091 = 0.0970657 loss) | |
| I0510 16:16:52.860013 10926 solver.cpp:245] Train net output #132: loss3/loss08 = 0.0364358 (* 0.0909091 = 0.00331235 loss) | |
| I0510 16:16:52.860028 10926 solver.cpp:245] Train net output #133: loss3/loss09 = 0.00286773 (* 0.0909091 = 0.000260703 loss) | |
| I0510 16:16:52.860044 10926 solver.cpp:245] Train net output #134: loss3/loss10 = 0.00241808 (* 0.0909091 = 0.000219825 loss) | |
| I0510 16:16:52.860059 10926 solver.cpp:245] Train net output #135: loss3/loss11 = 0.00120506 (* 0.0909091 = 0.000109551 loss) | |
| I0510 16:16:52.860074 10926 solver.cpp:245] Train net output #136: loss3/loss12 = 0.0013025 (* 0.0909091 = 0.000118409 loss) | |
| I0510 16:16:52.860088 10926 solver.cpp:245] Train net output #137: loss3/loss13 = 0.0012381 (* 0.0909091 = 0.000112554 loss) | |
| I0510 16:16:52.860103 10926 solver.cpp:245] Train net output #138: loss3/loss14 = 0.0012399 (* 0.0909091 = 0.000112718 loss) | |
| I0510 16:16:52.860117 10926 solver.cpp:245] Train net output #139: loss3/loss15 = 0.000646559 (* 0.0909091 = 5.87781e-05 loss) | |
| I0510 16:16:52.860131 10926 solver.cpp:245] Train net output #140: loss3/loss16 = 0.000656192 (* 0.0909091 = 5.96538e-05 loss) | |
| I0510 16:16:52.860146 10926 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00092287 (* 0.0909091 = 8.38973e-05 loss) | |
| I0510 16:16:52.860159 10926 solver.cpp:245] Train net output #142: loss3/loss18 = 0.000591754 (* 0.0909091 = 5.37958e-05 loss) | |
| I0510 16:16:52.860174 10926 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00057204 (* 0.0909091 = 5.20037e-05 loss) | |
| I0510 16:16:52.860188 10926 solver.cpp:245] Train net output #144: loss3/loss20 = 0.000327915 (* 0.0909091 = 2.98105e-05 loss) | |
| I0510 16:16:52.860203 10926 solver.cpp:245] Train net output #145: loss3/loss21 = 0.000289609 (* 0.0909091 = 2.63281e-05 loss) | |
| I0510 16:16:52.860216 10926 solver.cpp:245] Train net output #146: loss3/loss22 = 0.000287532 (* 0.0909091 = 2.61393e-05 loss) | |
| I0510 16:16:52.860239 10926 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0510 16:16:52.860254 10926 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0510 16:16:52.860265 10926 solver.cpp:245] Train net output #149: total_confidence = 8.20407e-05 | |
| I0510 16:16:52.860277 10926 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.000222245 | |
| I0510 16:16:52.860291 10926 sgd_solver.cpp:106] Iteration 10000, lr = 0.001 | |
| I0510 16:17:10.619710 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 33.7977 > 30) by scale factor 0.887635 | |
| I0510 16:17:21.243649 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 32.3301 > 30) by scale factor 0.927928 | |
| I0510 16:18:31.394407 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 66.6286 > 30) by scale factor 0.450257 | |
| I0510 16:19:14.613050 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 31.6327 > 30) by scale factor 0.948385 | |
| I0510 16:19:20.093835 10926 solver.cpp:229] Iteration 10500, loss = 10.4732 | |
| I0510 16:19:20.093895 10926 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0222222 | |
| I0510 16:19:20.093914 10926 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.125 | |
| I0510 16:19:20.093930 10926 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0510 16:19:20.093942 10926 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0510 16:19:20.093955 10926 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0 | |
| I0510 16:19:20.093967 10926 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.375 | |
| I0510 16:19:20.093981 10926 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.625 | |
| I0510 16:19:20.093994 10926 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.75 | |
| I0510 16:19:20.094007 10926 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.75 | |
| I0510 16:19:20.094020 10926 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.875 | |
| I0510 16:19:20.094033 10926 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.875 | |
| I0510 16:19:20.094046 10926 solver.cpp:245] Train net output #11: loss1/accuracy11 = 0.875 | |
| I0510 16:19:20.094058 10926 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0510 16:19:20.094072 10926 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0510 16:19:20.094084 10926 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0510 16:19:20.094096 10926 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0510 16:19:20.094108 10926 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0510 16:19:20.094120 10926 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0510 16:19:20.094132 10926 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0510 16:19:20.094147 10926 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0510 16:19:20.094161 10926 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0510 16:19:20.094172 10926 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0510 16:19:20.094184 10926 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0510 16:19:20.094197 10926 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.732955 | |
| I0510 16:19:20.094208 10926 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.111111 | |
| I0510 16:19:20.094224 10926 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.72029 (* 0.3 = 1.11609 loss) | |
| I0510 16:19:20.094240 10926 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.13318 (* 0.3 = 0.339953 loss) | |
| I0510 16:19:20.094254 10926 solver.cpp:245] Train net output #27: loss1/loss01 = 3.80729 (* 0.0272727 = 0.103835 loss) | |
| I0510 16:19:20.094269 10926 solver.cpp:245] Train net output #28: loss1/loss02 = 3.7194 (* 0.0272727 = 0.101438 loss) | |
| I0510 16:19:20.094283 10926 solver.cpp:245] Train net output #29: loss1/loss03 = 4.27357 (* 0.0272727 = 0.116552 loss) | |
| I0510 16:19:20.094297 10926 solver.cpp:245] Train net output #30: loss1/loss04 = 3.96819 (* 0.0272727 = 0.108223 loss) | |
| I0510 16:19:20.094312 10926 solver.cpp:245] Train net output #31: loss1/loss05 = 3.07584 (* 0.0272727 = 0.0838865 loss) | |
| I0510 16:19:20.094326 10926 solver.cpp:245] Train net output #32: loss1/loss06 = 2.13635 (* 0.0272727 = 0.0582641 loss) | |
| I0510 16:19:20.094341 10926 solver.cpp:245] Train net output #33: loss1/loss07 = 1.12058 (* 0.0272727 = 0.0305612 loss) | |
| I0510 16:19:20.094354 10926 solver.cpp:245] Train net output #34: loss1/loss08 = 1.07464 (* 0.0272727 = 0.0293083 loss) | |
| I0510 16:19:20.094368 10926 solver.cpp:245] Train net output #35: loss1/loss09 = 0.725887 (* 0.0272727 = 0.0197969 loss) | |
| I0510 16:19:20.094383 10926 solver.cpp:245] Train net output #36: loss1/loss10 = 0.505362 (* 0.0272727 = 0.0137826 loss) | |
| I0510 16:19:20.094396 10926 solver.cpp:245] Train net output #37: loss1/loss11 = 0.686158 (* 0.0272727 = 0.0187134 loss) | |
| I0510 16:19:20.094410 10926 solver.cpp:245] Train net output #38: loss1/loss12 = 0.1229 (* 0.0272727 = 0.00335182 loss) | |
| I0510 16:19:20.094455 10926 solver.cpp:245] Train net output #39: loss1/loss13 = 0.0627518 (* 0.0272727 = 0.00171141 loss) | |
| I0510 16:19:20.094471 10926 solver.cpp:245] Train net output #40: loss1/loss14 = 0.063495 (* 0.0272727 = 0.00173168 loss) | |
| I0510 16:19:20.094486 10926 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0851151 (* 0.0272727 = 0.00232132 loss) | |
| I0510 16:19:20.094501 10926 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0427971 (* 0.0272727 = 0.00116719 loss) | |
| I0510 16:19:20.094516 10926 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0164593 (* 0.0272727 = 0.000448889 loss) | |
| I0510 16:19:20.094529 10926 solver.cpp:245] Train net output #44: loss1/loss18 = 0.00976868 (* 0.0272727 = 0.000266419 loss) | |
| I0510 16:19:20.094543 10926 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0128834 (* 0.0272727 = 0.000351366 loss) | |
| I0510 16:19:20.094558 10926 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0139284 (* 0.0272727 = 0.000379864 loss) | |
| I0510 16:19:20.094573 10926 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00914977 (* 0.0272727 = 0.000249539 loss) | |
| I0510 16:19:20.094588 10926 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0144791 (* 0.0272727 = 0.000394884 loss) | |
| I0510 16:19:20.094599 10926 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0444444 | |
| I0510 16:19:20.094612 10926 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.125 | |
| I0510 16:19:20.094624 10926 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0510 16:19:20.094636 10926 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0510 16:19:20.094648 10926 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0 | |
| I0510 16:19:20.094660 10926 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.375 | |
| I0510 16:19:20.094671 10926 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.75 | |
| I0510 16:19:20.094683 10926 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.75 | |
| I0510 16:19:20.094696 10926 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.75 | |
| I0510 16:19:20.094710 10926 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.875 | |
| I0510 16:19:20.094722 10926 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.875 | |
| I0510 16:19:20.094735 10926 solver.cpp:245] Train net output #60: loss2/accuracy11 = 0.875 | |
| I0510 16:19:20.094748 10926 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0510 16:19:20.094759 10926 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0510 16:19:20.094771 10926 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0510 16:19:20.094782 10926 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0510 16:19:20.094794 10926 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0510 16:19:20.094806 10926 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0510 16:19:20.094817 10926 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0510 16:19:20.094830 10926 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0510 16:19:20.094841 10926 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0510 16:19:20.094852 10926 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0510 16:19:20.094864 10926 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0510 16:19:20.094876 10926 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.727273 | |
| I0510 16:19:20.094887 10926 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.155556 | |
| I0510 16:19:20.094902 10926 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.55294 (* 0.3 = 1.06588 loss) | |
| I0510 16:19:20.094915 10926 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.19879 (* 0.3 = 0.359636 loss) | |
| I0510 16:19:20.094930 10926 solver.cpp:245] Train net output #76: loss2/loss01 = 3.79512 (* 0.0272727 = 0.103503 loss) | |
| I0510 16:19:20.094944 10926 solver.cpp:245] Train net output #77: loss2/loss02 = 3.67865 (* 0.0272727 = 0.100327 loss) | |
| I0510 16:19:20.094970 10926 solver.cpp:245] Train net output #78: loss2/loss03 = 4.20255 (* 0.0272727 = 0.114615 loss) | |
| I0510 16:19:20.094986 10926 solver.cpp:245] Train net output #79: loss2/loss04 = 4.44986 (* 0.0272727 = 0.12136 loss) | |
| I0510 16:19:20.095000 10926 solver.cpp:245] Train net output #80: loss2/loss05 = 2.80463 (* 0.0272727 = 0.0764899 loss) | |
| I0510 16:19:20.095016 10926 solver.cpp:245] Train net output #81: loss2/loss06 = 1.87973 (* 0.0272727 = 0.0512653 loss) | |
| I0510 16:19:20.095029 10926 solver.cpp:245] Train net output #82: loss2/loss07 = 1.10654 (* 0.0272727 = 0.0301783 loss) | |
| I0510 16:19:20.095043 10926 solver.cpp:245] Train net output #83: loss2/loss08 = 1.28158 (* 0.0272727 = 0.0349521 loss) | |
| I0510 16:19:20.095057 10926 solver.cpp:245] Train net output #84: loss2/loss09 = 0.571877 (* 0.0272727 = 0.0155966 loss) | |
| I0510 16:19:20.095072 10926 solver.cpp:245] Train net output #85: loss2/loss10 = 0.472077 (* 0.0272727 = 0.0128748 loss) | |
| I0510 16:19:20.095085 10926 solver.cpp:245] Train net output #86: loss2/loss11 = 0.722579 (* 0.0272727 = 0.0197067 loss) | |
| I0510 16:19:20.095099 10926 solver.cpp:245] Train net output #87: loss2/loss12 = 0.0905124 (* 0.0272727 = 0.00246852 loss) | |
| I0510 16:19:20.095114 10926 solver.cpp:245] Train net output #88: loss2/loss13 = 0.0665277 (* 0.0272727 = 0.00181439 loss) | |
| I0510 16:19:20.095129 10926 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0265542 (* 0.0272727 = 0.000724205 loss) | |
| I0510 16:19:20.095142 10926 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0387786 (* 0.0272727 = 0.0010576 loss) | |
| I0510 16:19:20.095156 10926 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0171798 (* 0.0272727 = 0.00046854 loss) | |
| I0510 16:19:20.095170 10926 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0122821 (* 0.0272727 = 0.000334966 loss) | |
| I0510 16:19:20.095185 10926 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0196053 (* 0.0272727 = 0.00053469 loss) | |
| I0510 16:19:20.095202 10926 solver.cpp:245] Train net output #94: loss2/loss19 = 0.0197399 (* 0.0272727 = 0.00053836 loss) | |
| I0510 16:19:20.095217 10926 solver.cpp:245] Train net output #95: loss2/loss20 = 0.0120674 (* 0.0272727 = 0.00032911 loss) | |
| I0510 16:19:20.095232 10926 solver.cpp:245] Train net output #96: loss2/loss21 = 0.0231698 (* 0.0272727 = 0.000631903 loss) | |
| I0510 16:19:20.095245 10926 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0293305 (* 0.0272727 = 0.000799922 loss) | |
| I0510 16:19:20.095257 10926 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0666667 | |
| I0510 16:19:20.095270 10926 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0510 16:19:20.095283 10926 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0510 16:19:20.095293 10926 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0510 16:19:20.095305 10926 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0 | |
| I0510 16:19:20.095316 10926 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.375 | |
| I0510 16:19:20.095329 10926 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.625 | |
| I0510 16:19:20.095340 10926 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.75 | |
| I0510 16:19:20.095352 10926 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.75 | |
| I0510 16:19:20.095365 10926 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.875 | |
| I0510 16:19:20.095376 10926 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.875 | |
| I0510 16:19:20.095388 10926 solver.cpp:245] Train net output #109: loss3/accuracy11 = 0.875 | |
| I0510 16:19:20.095401 10926 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0510 16:19:20.095412 10926 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0510 16:19:20.095424 10926 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0510 16:19:20.095435 10926 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0510 16:19:20.095458 10926 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0510 16:19:20.095471 10926 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0510 16:19:20.095484 10926 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0510 16:19:20.095494 10926 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0510 16:19:20.095506 10926 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0510 16:19:20.095518 10926 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0510 16:19:20.095530 10926 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0510 16:19:20.095541 10926 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.744318 | |
| I0510 16:19:20.095553 10926 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.133333 | |
| I0510 16:19:20.095567 10926 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.62425 (* 1 = 3.62425 loss) | |
| I0510 16:19:20.095582 10926 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.09837 (* 1 = 1.09837 loss) | |
| I0510 16:19:20.095595 10926 solver.cpp:245] Train net output #125: loss3/loss01 = 4.01626 (* 0.0909091 = 0.365115 loss) | |
| I0510 16:19:20.095609 10926 solver.cpp:245] Train net output #126: loss3/loss02 = 4.06199 (* 0.0909091 = 0.369272 loss) | |
| I0510 16:19:20.095620 10926 solver.cpp:245] Train net output #127: loss3/loss03 = 4.06474 (* 0.0909091 = 0.369522 loss) | |
| I0510 16:19:20.095629 10926 solver.cpp:245] Train net output #128: loss3/loss04 = 4.28649 (* 0.0909091 = 0.389681 loss) | |
| I0510 16:19:20.095644 10926 solver.cpp:245] Train net output #129: loss3/loss05 = 2.58566 (* 0.0909091 = 0.23506 loss) | |
| I0510 16:19:20.095659 10926 solver.cpp:245] Train net output #130: loss3/loss06 = 1.60853 (* 0.0909091 = 0.14623 loss) | |
| I0510 16:19:20.095672 10926 solver.cpp:245] Train net output #131: loss3/loss07 = 0.797122 (* 0.0909091 = 0.0724656 loss) | |
| I0510 16:19:20.095687 10926 solver.cpp:245] Train net output #132: loss3/loss08 = 1.20045 (* 0.0909091 = 0.109132 loss) | |
| I0510 16:19:20.095701 10926 solver.cpp:245] Train net output #133: loss3/loss09 = 0.710642 (* 0.0909091 = 0.0646039 loss) | |
| I0510 16:19:20.095715 10926 solver.cpp:245] Train net output #134: loss3/loss10 = 0.362958 (* 0.0909091 = 0.0329962 loss) | |
| I0510 16:19:20.095729 10926 solver.cpp:245] Train net output #135: loss3/loss11 = 0.698595 (* 0.0909091 = 0.0635087 loss) | |
| I0510 16:19:20.095744 10926 solver.cpp:245] Train net output #136: loss3/loss12 = 0.0527347 (* 0.0909091 = 0.00479407 loss) | |
| I0510 16:19:20.095762 10926 solver.cpp:245] Train net output #137: loss3/loss13 = 0.0642719 (* 0.0909091 = 0.0058429 loss) | |
| I0510 16:19:20.095777 10926 solver.cpp:245] Train net output #138: loss3/loss14 = 0.0313785 (* 0.0909091 = 0.00285259 loss) | |
| I0510 16:19:20.095791 10926 solver.cpp:245] Train net output #139: loss3/loss15 = 0.0247364 (* 0.0909091 = 0.00224876 loss) | |
| I0510 16:19:20.095804 10926 solver.cpp:245] Train net output #140: loss3/loss16 = 0.0127571 (* 0.0909091 = 0.00115974 loss) | |
| I0510 16:19:20.095819 10926 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00759903 (* 0.0909091 = 0.000690821 loss) | |
| I0510 16:19:20.095832 10926 solver.cpp:245] Train net output #142: loss3/loss18 = 0.0057635 (* 0.0909091 = 0.000523955 loss) | |
| I0510 16:19:20.095846 10926 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00294479 (* 0.0909091 = 0.000267709 loss) | |
| I0510 16:19:20.095860 10926 solver.cpp:245] Train net output #144: loss3/loss20 = 0.0022353 (* 0.0909091 = 0.000203209 loss) | |
| I0510 16:19:20.095875 10926 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00268526 (* 0.0909091 = 0.000244115 loss) | |
| I0510 16:19:20.095888 10926 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00203738 (* 0.0909091 = 0.000185216 loss) | |
| I0510 16:19:20.095901 10926 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0510 16:19:20.095912 10926 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0510 16:19:20.095934 10926 solver.cpp:245] Train net output #149: total_confidence = 9.79613e-05 | |
| I0510 16:19:20.095948 10926 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.000532045 | |
| I0510 16:19:20.095963 10926 sgd_solver.cpp:106] Iteration 10500, lr = 0.001 | |
| I0510 16:19:37.821148 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 30.4463 > 30) by scale factor 0.985342 | |
| I0510 16:19:44.913450 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 57.9982 > 30) by scale factor 0.517258 | |
| I0510 16:20:14.211292 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 42.7717 > 30) by scale factor 0.701398 | |
| I0510 16:20:37.816396 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 30.2887 > 30) by scale factor 0.990468 | |
| I0510 16:21:47.731353 10926 solver.cpp:229] Iteration 11000, loss = 10.3039 | |
| I0510 16:21:47.731501 10926 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0208333 | |
| I0510 16:21:47.731524 10926 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.125 | |
| I0510 16:21:47.731539 10926 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0510 16:21:47.731552 10926 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0510 16:21:47.731565 10926 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.125 | |
| I0510 16:21:47.731578 10926 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.375 | |
| I0510 16:21:47.731591 10926 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.5 | |
| I0510 16:21:47.731612 10926 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.625 | |
| I0510 16:21:47.731626 10926 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.625 | |
| I0510 16:21:47.731637 10926 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.875 | |
| I0510 16:21:47.731652 10926 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.875 | |
| I0510 16:21:47.731664 10926 solver.cpp:245] Train net output #11: loss1/accuracy11 = 0.875 | |
| I0510 16:21:47.731684 10926 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0510 16:21:47.731698 10926 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0510 16:21:47.731709 10926 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0510 16:21:47.731721 10926 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0510 16:21:47.731734 10926 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0510 16:21:47.731745 10926 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0510 16:21:47.731756 10926 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0510 16:21:47.731768 10926 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0510 16:21:47.731781 10926 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0510 16:21:47.731792 10926 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0510 16:21:47.731804 10926 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0510 16:21:47.731815 10926 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.715909 | |
| I0510 16:21:47.731827 10926 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.208333 | |
| I0510 16:21:47.731845 10926 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.33149 (* 0.3 = 0.999446 loss) | |
| I0510 16:21:47.731860 10926 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.11303 (* 0.3 = 0.333908 loss) | |
| I0510 16:21:47.731879 10926 solver.cpp:245] Train net output #27: loss1/loss01 = 3.30484 (* 0.0272727 = 0.090132 loss) | |
| I0510 16:21:47.731894 10926 solver.cpp:245] Train net output #28: loss1/loss02 = 4.01853 (* 0.0272727 = 0.109596 loss) | |
| I0510 16:21:47.731909 10926 solver.cpp:245] Train net output #29: loss1/loss03 = 3.45 (* 0.0272727 = 0.094091 loss) | |
| I0510 16:21:47.731930 10926 solver.cpp:245] Train net output #30: loss1/loss04 = 3.41139 (* 0.0272727 = 0.0930379 loss) | |
| I0510 16:21:47.731945 10926 solver.cpp:245] Train net output #31: loss1/loss05 = 2.64962 (* 0.0272727 = 0.0722625 loss) | |
| I0510 16:21:47.731958 10926 solver.cpp:245] Train net output #32: loss1/loss06 = 2.35206 (* 0.0272727 = 0.064147 loss) | |
| I0510 16:21:47.731972 10926 solver.cpp:245] Train net output #33: loss1/loss07 = 1.55151 (* 0.0272727 = 0.0423138 loss) | |
| I0510 16:21:47.731994 10926 solver.cpp:245] Train net output #34: loss1/loss08 = 1.85499 (* 0.0272727 = 0.0505905 loss) | |
| I0510 16:21:47.732008 10926 solver.cpp:245] Train net output #35: loss1/loss09 = 0.623977 (* 0.0272727 = 0.0170176 loss) | |
| I0510 16:21:47.732023 10926 solver.cpp:245] Train net output #36: loss1/loss10 = 0.516777 (* 0.0272727 = 0.0140939 loss) | |
| I0510 16:21:47.732038 10926 solver.cpp:245] Train net output #37: loss1/loss11 = 0.70463 (* 0.0272727 = 0.0192172 loss) | |
| I0510 16:21:47.732053 10926 solver.cpp:245] Train net output #38: loss1/loss12 = 0.0905421 (* 0.0272727 = 0.00246933 loss) | |
| I0510 16:21:47.732069 10926 solver.cpp:245] Train net output #39: loss1/loss13 = 0.0746304 (* 0.0272727 = 0.00203538 loss) | |
| I0510 16:21:47.732103 10926 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0988507 (* 0.0272727 = 0.00269593 loss) | |
| I0510 16:21:47.732120 10926 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0261129 (* 0.0272727 = 0.000712171 loss) | |
| I0510 16:21:47.732134 10926 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0134808 (* 0.0272727 = 0.000367658 loss) | |
| I0510 16:21:47.732149 10926 solver.cpp:245] Train net output #43: loss1/loss17 = 0.00248201 (* 0.0272727 = 6.76911e-05 loss) | |
| I0510 16:21:47.732163 10926 solver.cpp:245] Train net output #44: loss1/loss18 = 0.00301017 (* 0.0272727 = 8.20954e-05 loss) | |
| I0510 16:21:47.732177 10926 solver.cpp:245] Train net output #45: loss1/loss19 = 0.00213347 (* 0.0272727 = 5.81855e-05 loss) | |
| I0510 16:21:47.732192 10926 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00384644 (* 0.0272727 = 0.000104903 loss) | |
| I0510 16:21:47.732208 10926 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00339006 (* 0.0272727 = 9.24562e-05 loss) | |
| I0510 16:21:47.732221 10926 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0025041 (* 0.0272727 = 6.82938e-05 loss) | |
| I0510 16:21:47.732234 10926 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0416667 | |
| I0510 16:21:47.732246 10926 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0510 16:21:47.732259 10926 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0510 16:21:47.732270 10926 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0.125 | |
| I0510 16:21:47.732282 10926 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.25 | |
| I0510 16:21:47.732295 10926 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.375 | |
| I0510 16:21:47.732306 10926 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.5 | |
| I0510 16:21:47.732319 10926 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.625 | |
| I0510 16:21:47.732331 10926 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.625 | |
| I0510 16:21:47.732343 10926 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.875 | |
| I0510 16:21:47.732355 10926 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.875 | |
| I0510 16:21:47.732367 10926 solver.cpp:245] Train net output #60: loss2/accuracy11 = 0.875 | |
| I0510 16:21:47.732379 10926 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0510 16:21:47.732390 10926 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0510 16:21:47.732403 10926 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0510 16:21:47.732414 10926 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0510 16:21:47.732425 10926 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0510 16:21:47.732437 10926 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0510 16:21:47.732448 10926 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0510 16:21:47.732460 10926 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0510 16:21:47.732472 10926 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0510 16:21:47.732484 10926 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0510 16:21:47.732496 10926 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0510 16:21:47.732508 10926 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.727273 | |
| I0510 16:21:47.732524 10926 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.208333 | |
| I0510 16:21:47.732538 10926 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.20395 (* 0.3 = 0.961184 loss) | |
| I0510 16:21:47.732553 10926 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.04423 (* 0.3 = 0.31327 loss) | |
| I0510 16:21:47.732568 10926 solver.cpp:245] Train net output #76: loss2/loss01 = 3.9928 (* 0.0272727 = 0.108895 loss) | |
| I0510 16:21:47.732588 10926 solver.cpp:245] Train net output #77: loss2/loss02 = 3.72452 (* 0.0272727 = 0.101578 loss) | |
| I0510 16:21:47.732612 10926 solver.cpp:245] Train net output #78: loss2/loss03 = 3.28186 (* 0.0272727 = 0.0895052 loss) | |
| I0510 16:21:47.732628 10926 solver.cpp:245] Train net output #79: loss2/loss04 = 3.28995 (* 0.0272727 = 0.0897258 loss) | |
| I0510 16:21:47.732643 10926 solver.cpp:245] Train net output #80: loss2/loss05 = 2.33232 (* 0.0272727 = 0.0636087 loss) | |
| I0510 16:21:47.732666 10926 solver.cpp:245] Train net output #81: loss2/loss06 = 2.37936 (* 0.0272727 = 0.0648917 loss) | |
| I0510 16:21:47.732679 10926 solver.cpp:245] Train net output #82: loss2/loss07 = 1.62047 (* 0.0272727 = 0.0441945 loss) | |
| I0510 16:21:47.732693 10926 solver.cpp:245] Train net output #83: loss2/loss08 = 1.71418 (* 0.0272727 = 0.0467503 loss) | |
| I0510 16:21:47.732707 10926 solver.cpp:245] Train net output #84: loss2/loss09 = 0.530798 (* 0.0272727 = 0.0144763 loss) | |
| I0510 16:21:47.732722 10926 solver.cpp:245] Train net output #85: loss2/loss10 = 0.40209 (* 0.0272727 = 0.0109661 loss) | |
| I0510 16:21:47.732735 10926 solver.cpp:245] Train net output #86: loss2/loss11 = 0.457517 (* 0.0272727 = 0.0124777 loss) | |
| I0510 16:21:47.732750 10926 solver.cpp:245] Train net output #87: loss2/loss12 = 0.220527 (* 0.0272727 = 0.00601437 loss) | |
| I0510 16:21:47.732765 10926 solver.cpp:245] Train net output #88: loss2/loss13 = 0.136076 (* 0.0272727 = 0.00371115 loss) | |
| I0510 16:21:47.732779 10926 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0622378 (* 0.0272727 = 0.00169739 loss) | |
| I0510 16:21:47.732794 10926 solver.cpp:245] Train net output #90: loss2/loss15 = 0.127686 (* 0.0272727 = 0.00348233 loss) | |
| I0510 16:21:47.732808 10926 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0343945 (* 0.0272727 = 0.000938033 loss) | |
| I0510 16:21:47.732822 10926 solver.cpp:245] Train net output #92: loss2/loss17 = 0.00962849 (* 0.0272727 = 0.000262595 loss) | |
| I0510 16:21:47.732836 10926 solver.cpp:245] Train net output #93: loss2/loss18 = 0.00422987 (* 0.0272727 = 0.00011536 loss) | |
| I0510 16:21:47.732851 10926 solver.cpp:245] Train net output #94: loss2/loss19 = 0.00542832 (* 0.0272727 = 0.000148045 loss) | |
| I0510 16:21:47.732866 10926 solver.cpp:245] Train net output #95: loss2/loss20 = 0.0241496 (* 0.0272727 = 0.000658627 loss) | |
| I0510 16:21:47.732879 10926 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00693993 (* 0.0272727 = 0.000189271 loss) | |
| I0510 16:21:47.732893 10926 solver.cpp:245] Train net output #97: loss2/loss22 = 0.008154 (* 0.0272727 = 0.000222382 loss) | |
| I0510 16:21:47.732906 10926 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0625 | |
| I0510 16:21:47.732919 10926 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0510 16:21:47.732934 10926 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.125 | |
| I0510 16:21:47.732947 10926 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0510 16:21:47.732959 10926 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.375 | |
| I0510 16:21:47.732971 10926 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.5 | |
| I0510 16:21:47.732983 10926 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.5 | |
| I0510 16:21:47.732995 10926 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.625 | |
| I0510 16:21:47.733007 10926 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.75 | |
| I0510 16:21:47.733019 10926 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.875 | |
| I0510 16:21:47.733031 10926 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.875 | |
| I0510 16:21:47.733042 10926 solver.cpp:245] Train net output #109: loss3/accuracy11 = 0.875 | |
| I0510 16:21:47.733054 10926 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0510 16:21:47.733067 10926 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0510 16:21:47.733078 10926 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0510 16:21:47.733089 10926 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0510 16:21:47.733101 10926 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0510 16:21:47.733139 10926 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0510 16:21:47.733155 10926 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0510 16:21:47.733167 10926 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0510 16:21:47.733180 10926 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0510 16:21:47.733191 10926 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0510 16:21:47.733201 10926 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0510 16:21:47.733208 10926 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.732955 | |
| I0510 16:21:47.733222 10926 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.25 | |
| I0510 16:21:47.733237 10926 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.19564 (* 1 = 3.19564 loss) | |
| I0510 16:21:47.733252 10926 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.00247 (* 1 = 1.00247 loss) | |
| I0510 16:21:47.733265 10926 solver.cpp:245] Train net output #125: loss3/loss01 = 3.22457 (* 0.0909091 = 0.293143 loss) | |
| I0510 16:21:47.733279 10926 solver.cpp:245] Train net output #126: loss3/loss02 = 3.03152 (* 0.0909091 = 0.275593 loss) | |
| I0510 16:21:47.733294 10926 solver.cpp:245] Train net output #127: loss3/loss03 = 3.01913 (* 0.0909091 = 0.274466 loss) | |
| I0510 16:21:47.733307 10926 solver.cpp:245] Train net output #128: loss3/loss04 = 2.96561 (* 0.0909091 = 0.269601 loss) | |
| I0510 16:21:47.733322 10926 solver.cpp:245] Train net output #129: loss3/loss05 = 2.31555 (* 0.0909091 = 0.210505 loss) | |
| I0510 16:21:47.733336 10926 solver.cpp:245] Train net output #130: loss3/loss06 = 1.85645 (* 0.0909091 = 0.168768 loss) | |
| I0510 16:21:47.733350 10926 solver.cpp:245] Train net output #131: loss3/loss07 = 1.45505 (* 0.0909091 = 0.132278 loss) | |
| I0510 16:21:47.733364 10926 solver.cpp:245] Train net output #132: loss3/loss08 = 1.40058 (* 0.0909091 = 0.127326 loss) | |
| I0510 16:21:47.733379 10926 solver.cpp:245] Train net output #133: loss3/loss09 = 0.336237 (* 0.0909091 = 0.030567 loss) | |
| I0510 16:21:47.733393 10926 solver.cpp:245] Train net output #134: loss3/loss10 = 0.395846 (* 0.0909091 = 0.035986 loss) | |
| I0510 16:21:47.733407 10926 solver.cpp:245] Train net output #135: loss3/loss11 = 0.499674 (* 0.0909091 = 0.0454249 loss) | |
| I0510 16:21:47.733422 10926 solver.cpp:245] Train net output #136: loss3/loss12 = 0.2047 (* 0.0909091 = 0.0186091 loss) | |
| I0510 16:21:47.733435 10926 solver.cpp:245] Train net output #137: loss3/loss13 = 0.186312 (* 0.0909091 = 0.0169375 loss) | |
| I0510 16:21:47.733449 10926 solver.cpp:245] Train net output #138: loss3/loss14 = 0.126297 (* 0.0909091 = 0.0114816 loss) | |
| I0510 16:21:47.733464 10926 solver.cpp:245] Train net output #139: loss3/loss15 = 0.0719167 (* 0.0909091 = 0.00653789 loss) | |
| I0510 16:21:47.733479 10926 solver.cpp:245] Train net output #140: loss3/loss16 = 0.040479 (* 0.0909091 = 0.00367991 loss) | |
| I0510 16:21:47.733492 10926 solver.cpp:245] Train net output #141: loss3/loss17 = 0.0258715 (* 0.0909091 = 0.00235195 loss) | |
| I0510 16:21:47.733506 10926 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00601623 (* 0.0909091 = 0.00054693 loss) | |
| I0510 16:21:47.733520 10926 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00366517 (* 0.0909091 = 0.000333198 loss) | |
| I0510 16:21:47.733535 10926 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00309665 (* 0.0909091 = 0.000281514 loss) | |
| I0510 16:21:47.733548 10926 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00114106 (* 0.0909091 = 0.000103733 loss) | |
| I0510 16:21:47.733566 10926 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00130149 (* 0.0909091 = 0.000118317 loss) | |
| I0510 16:21:47.733578 10926 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0510 16:21:47.733590 10926 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0510 16:21:47.733603 10926 solver.cpp:245] Train net output #149: total_confidence = 9.47509e-05 | |
| I0510 16:21:47.733631 10926 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.000392119 | |
| I0510 16:21:47.733646 10926 sgd_solver.cpp:106] Iteration 11000, lr = 0.001 | |
| I0510 16:22:09.954046 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 43.3483 > 30) by scale factor 0.692068 | |
| I0510 16:22:21.158357 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 38.4295 > 30) by scale factor 0.78065 | |
| I0510 16:22:24.412587 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 44.4735 > 30) by scale factor 0.674559 | |
| I0510 16:22:35.340252 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 36.2201 > 30) by scale factor 0.82827 | |
| I0510 16:23:16.180364 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 31.9989 > 30) by scale factor 0.937532 | |
| I0510 16:23:47.926141 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 34.9486 > 30) by scale factor 0.858403 | |
| I0510 16:23:50.862079 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 42.8695 > 30) by scale factor 0.699798 | |
| I0510 16:24:14.841976 10926 solver.cpp:229] Iteration 11500, loss = 10.3057 | |
| I0510 16:24:14.842036 10926 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0769231 | |
| I0510 16:24:14.842056 10926 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.125 | |
| I0510 16:24:14.842069 10926 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0510 16:24:14.842082 10926 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0510 16:24:14.842094 10926 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.125 | |
| I0510 16:24:14.842108 10926 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.125 | |
| I0510 16:24:14.842120 10926 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.125 | |
| I0510 16:24:14.842133 10926 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.25 | |
| I0510 16:24:14.842147 10926 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.375 | |
| I0510 16:24:14.842159 10926 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.75 | |
| I0510 16:24:14.842172 10926 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.75 | |
| I0510 16:24:14.842186 10926 solver.cpp:245] Train net output #11: loss1/accuracy11 = 0.875 | |
| I0510 16:24:14.842200 10926 solver.cpp:245] Train net output #12: loss1/accuracy12 = 0.875 | |
| I0510 16:24:14.842221 10926 solver.cpp:245] Train net output #13: loss1/accuracy13 = 0.875 | |
| I0510 16:24:14.842239 10926 solver.cpp:245] Train net output #14: loss1/accuracy14 = 0.875 | |
| I0510 16:24:14.842252 10926 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0510 16:24:14.842264 10926 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0510 16:24:14.842277 10926 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0510 16:24:14.842288 10926 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0510 16:24:14.842300 10926 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0510 16:24:14.842313 10926 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0510 16:24:14.842324 10926 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0510 16:24:14.842336 10926 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0510 16:24:14.842347 10926 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.619318 | |
| I0510 16:24:14.842360 10926 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.230769 | |
| I0510 16:24:14.842376 10926 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.24445 (* 0.3 = 0.973336 loss) | |
| I0510 16:24:14.842391 10926 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.42733 (* 0.3 = 0.428198 loss) | |
| I0510 16:24:14.842406 10926 solver.cpp:245] Train net output #27: loss1/loss01 = 3.41698 (* 0.0272727 = 0.0931904 loss) | |
| I0510 16:24:14.842420 10926 solver.cpp:245] Train net output #28: loss1/loss02 = 3.37831 (* 0.0272727 = 0.0921357 loss) | |
| I0510 16:24:14.842434 10926 solver.cpp:245] Train net output #29: loss1/loss03 = 3.67287 (* 0.0272727 = 0.100169 loss) | |
| I0510 16:24:14.842449 10926 solver.cpp:245] Train net output #30: loss1/loss04 = 3.60233 (* 0.0272727 = 0.0982454 loss) | |
| I0510 16:24:14.842464 10926 solver.cpp:245] Train net output #31: loss1/loss05 = 3.02614 (* 0.0272727 = 0.082531 loss) | |
| I0510 16:24:14.842478 10926 solver.cpp:245] Train net output #32: loss1/loss06 = 3.56041 (* 0.0272727 = 0.097102 loss) | |
| I0510 16:24:14.842499 10926 solver.cpp:245] Train net output #33: loss1/loss07 = 3.39003 (* 0.0272727 = 0.0924554 loss) | |
| I0510 16:24:14.842517 10926 solver.cpp:245] Train net output #34: loss1/loss08 = 3.06555 (* 0.0272727 = 0.0836059 loss) | |
| I0510 16:24:14.842532 10926 solver.cpp:245] Train net output #35: loss1/loss09 = 1.40624 (* 0.0272727 = 0.0383521 loss) | |
| I0510 16:24:14.842547 10926 solver.cpp:245] Train net output #36: loss1/loss10 = 1.4275 (* 0.0272727 = 0.0389317 loss) | |
| I0510 16:24:14.842561 10926 solver.cpp:245] Train net output #37: loss1/loss11 = 0.578228 (* 0.0272727 = 0.0157698 loss) | |
| I0510 16:24:14.842608 10926 solver.cpp:245] Train net output #38: loss1/loss12 = 0.864447 (* 0.0272727 = 0.0235758 loss) | |
| I0510 16:24:14.842624 10926 solver.cpp:245] Train net output #39: loss1/loss13 = 0.663043 (* 0.0272727 = 0.018083 loss) | |
| I0510 16:24:14.842638 10926 solver.cpp:245] Train net output #40: loss1/loss14 = 0.729721 (* 0.0272727 = 0.0199015 loss) | |
| I0510 16:24:14.842653 10926 solver.cpp:245] Train net output #41: loss1/loss15 = 0.18177 (* 0.0272727 = 0.00495737 loss) | |
| I0510 16:24:14.842669 10926 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0391338 (* 0.0272727 = 0.00106729 loss) | |
| I0510 16:24:14.842682 10926 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0308305 (* 0.0272727 = 0.000840831 loss) | |
| I0510 16:24:14.842700 10926 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0548242 (* 0.0272727 = 0.0014952 loss) | |
| I0510 16:24:14.842715 10926 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0101251 (* 0.0272727 = 0.000276139 loss) | |
| I0510 16:24:14.842730 10926 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0146954 (* 0.0272727 = 0.000400783 loss) | |
| I0510 16:24:14.842746 10926 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0223438 (* 0.0272727 = 0.000609376 loss) | |
| I0510 16:24:14.842759 10926 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0131038 (* 0.0272727 = 0.000357376 loss) | |
| I0510 16:24:14.842772 10926 solver.cpp:245] Train net output #49: loss2/accuracy = 0 | |
| I0510 16:24:14.842787 10926 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0510 16:24:14.842799 10926 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0.125 | |
| I0510 16:24:14.842811 10926 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0.125 | |
| I0510 16:24:14.842823 10926 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0 | |
| I0510 16:24:14.842835 10926 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.125 | |
| I0510 16:24:14.842847 10926 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.125 | |
| I0510 16:24:14.842859 10926 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.25 | |
| I0510 16:24:14.842870 10926 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.375 | |
| I0510 16:24:14.842883 10926 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.75 | |
| I0510 16:24:14.842895 10926 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.75 | |
| I0510 16:24:14.842907 10926 solver.cpp:245] Train net output #60: loss2/accuracy11 = 0.875 | |
| I0510 16:24:14.842919 10926 solver.cpp:245] Train net output #61: loss2/accuracy12 = 0.875 | |
| I0510 16:24:14.842931 10926 solver.cpp:245] Train net output #62: loss2/accuracy13 = 0.875 | |
| I0510 16:24:14.842943 10926 solver.cpp:245] Train net output #63: loss2/accuracy14 = 0.875 | |
| I0510 16:24:14.842955 10926 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0510 16:24:14.842967 10926 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0510 16:24:14.842979 10926 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0510 16:24:14.842990 10926 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0510 16:24:14.843003 10926 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0510 16:24:14.843014 10926 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0510 16:24:14.843025 10926 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0510 16:24:14.843036 10926 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0510 16:24:14.843049 10926 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.585227 | |
| I0510 16:24:14.843060 10926 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.184615 | |
| I0510 16:24:14.843073 10926 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.23045 (* 0.3 = 0.969134 loss) | |
| I0510 16:24:14.843088 10926 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.47918 (* 0.3 = 0.443755 loss) | |
| I0510 16:24:14.843102 10926 solver.cpp:245] Train net output #76: loss2/loss01 = 3.43835 (* 0.0272727 = 0.0937733 loss) | |
| I0510 16:24:14.843128 10926 solver.cpp:245] Train net output #77: loss2/loss02 = 3.29367 (* 0.0272727 = 0.0898272 loss) | |
| I0510 16:24:14.843144 10926 solver.cpp:245] Train net output #78: loss2/loss03 = 3.50557 (* 0.0272727 = 0.0956064 loss) | |
| I0510 16:24:14.843159 10926 solver.cpp:245] Train net output #79: loss2/loss04 = 3.94685 (* 0.0272727 = 0.107641 loss) | |
| I0510 16:24:14.843173 10926 solver.cpp:245] Train net output #80: loss2/loss05 = 3.32233 (* 0.0272727 = 0.090609 loss) | |
| I0510 16:24:14.843188 10926 solver.cpp:245] Train net output #81: loss2/loss06 = 3.31236 (* 0.0272727 = 0.0903371 loss) | |
| I0510 16:24:14.843202 10926 solver.cpp:245] Train net output #82: loss2/loss07 = 3.1043 (* 0.0272727 = 0.0846628 loss) | |
| I0510 16:24:14.843216 10926 solver.cpp:245] Train net output #83: loss2/loss08 = 3.60244 (* 0.0272727 = 0.0982482 loss) | |
| I0510 16:24:14.843230 10926 solver.cpp:245] Train net output #84: loss2/loss09 = 1.06618 (* 0.0272727 = 0.0290777 loss) | |
| I0510 16:24:14.843245 10926 solver.cpp:245] Train net output #85: loss2/loss10 = 1.16011 (* 0.0272727 = 0.0316392 loss) | |
| I0510 16:24:14.843260 10926 solver.cpp:245] Train net output #86: loss2/loss11 = 0.697417 (* 0.0272727 = 0.0190205 loss) | |
| I0510 16:24:14.843273 10926 solver.cpp:245] Train net output #87: loss2/loss12 = 0.492989 (* 0.0272727 = 0.0134451 loss) | |
| I0510 16:24:14.843287 10926 solver.cpp:245] Train net output #88: loss2/loss13 = 0.833736 (* 0.0272727 = 0.0227383 loss) | |
| I0510 16:24:14.843302 10926 solver.cpp:245] Train net output #89: loss2/loss14 = 0.461621 (* 0.0272727 = 0.0125897 loss) | |
| I0510 16:24:14.843315 10926 solver.cpp:245] Train net output #90: loss2/loss15 = 0.152208 (* 0.0272727 = 0.00415114 loss) | |
| I0510 16:24:14.843329 10926 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0756404 (* 0.0272727 = 0.00206292 loss) | |
| I0510 16:24:14.843343 10926 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0728589 (* 0.0272727 = 0.00198706 loss) | |
| I0510 16:24:14.843358 10926 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0333096 (* 0.0272727 = 0.000908443 loss) | |
| I0510 16:24:14.843372 10926 solver.cpp:245] Train net output #94: loss2/loss19 = 0.0258096 (* 0.0272727 = 0.000703898 loss) | |
| I0510 16:24:14.843386 10926 solver.cpp:245] Train net output #95: loss2/loss20 = 0.017162 (* 0.0272727 = 0.000468055 loss) | |
| I0510 16:24:14.843401 10926 solver.cpp:245] Train net output #96: loss2/loss21 = 0.033732 (* 0.0272727 = 0.000919963 loss) | |
| I0510 16:24:14.843420 10926 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0176556 (* 0.0272727 = 0.000481517 loss) | |
| I0510 16:24:14.843428 10926 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0923077 | |
| I0510 16:24:14.843443 10926 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.125 | |
| I0510 16:24:14.843464 10926 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0510 16:24:14.843477 10926 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0510 16:24:14.843490 10926 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0 | |
| I0510 16:24:14.843502 10926 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.125 | |
| I0510 16:24:14.843514 10926 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.125 | |
| I0510 16:24:14.843525 10926 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.375 | |
| I0510 16:24:14.843538 10926 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.375 | |
| I0510 16:24:14.843549 10926 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.625 | |
| I0510 16:24:14.843561 10926 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.75 | |
| I0510 16:24:14.843572 10926 solver.cpp:245] Train net output #109: loss3/accuracy11 = 0.875 | |
| I0510 16:24:14.843585 10926 solver.cpp:245] Train net output #110: loss3/accuracy12 = 0.875 | |
| I0510 16:24:14.843596 10926 solver.cpp:245] Train net output #111: loss3/accuracy13 = 0.875 | |
| I0510 16:24:14.843607 10926 solver.cpp:245] Train net output #112: loss3/accuracy14 = 0.875 | |
| I0510 16:24:14.843631 10926 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0510 16:24:14.843644 10926 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0510 16:24:14.843657 10926 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0510 16:24:14.843667 10926 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0510 16:24:14.843679 10926 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0510 16:24:14.843691 10926 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0510 16:24:14.843703 10926 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0510 16:24:14.843714 10926 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0510 16:24:14.843726 10926 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.642045 | |
| I0510 16:24:14.843737 10926 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.323077 | |
| I0510 16:24:14.843755 10926 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.04526 (* 1 = 3.04526 loss) | |
| I0510 16:24:14.843770 10926 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.31433 (* 1 = 1.31433 loss) | |
| I0510 16:24:14.843783 10926 solver.cpp:245] Train net output #125: loss3/loss01 = 2.84217 (* 0.0909091 = 0.258379 loss) | |
| I0510 16:24:14.843797 10926 solver.cpp:245] Train net output #126: loss3/loss02 = 2.87561 (* 0.0909091 = 0.261419 loss) | |
| I0510 16:24:14.843811 10926 solver.cpp:245] Train net output #127: loss3/loss03 = 3.19237 (* 0.0909091 = 0.290216 loss) | |
| I0510 16:24:14.843825 10926 solver.cpp:245] Train net output #128: loss3/loss04 = 3.60665 (* 0.0909091 = 0.327877 loss) | |
| I0510 16:24:14.843842 10926 solver.cpp:245] Train net output #129: loss3/loss05 = 3.03441 (* 0.0909091 = 0.275855 loss) | |
| I0510 16:24:14.843857 10926 solver.cpp:245] Train net output #130: loss3/loss06 = 3.16884 (* 0.0909091 = 0.288077 loss) | |
| I0510 16:24:14.843870 10926 solver.cpp:245] Train net output #131: loss3/loss07 = 2.30334 (* 0.0909091 = 0.209395 loss) | |
| I0510 16:24:14.843884 10926 solver.cpp:245] Train net output #132: loss3/loss08 = 2.86652 (* 0.0909091 = 0.260593 loss) | |
| I0510 16:24:14.843899 10926 solver.cpp:245] Train net output #133: loss3/loss09 = 1.13311 (* 0.0909091 = 0.10301 loss) | |
| I0510 16:24:14.843914 10926 solver.cpp:245] Train net output #134: loss3/loss10 = 0.950479 (* 0.0909091 = 0.0864072 loss) | |
| I0510 16:24:14.843927 10926 solver.cpp:245] Train net output #135: loss3/loss11 = 0.546292 (* 0.0909091 = 0.0496629 loss) | |
| I0510 16:24:14.843941 10926 solver.cpp:245] Train net output #136: loss3/loss12 = 0.560128 (* 0.0909091 = 0.0509207 loss) | |
| I0510 16:24:14.843955 10926 solver.cpp:245] Train net output #137: loss3/loss13 = 0.520819 (* 0.0909091 = 0.0473472 loss) | |
| I0510 16:24:14.843971 10926 solver.cpp:245] Train net output #138: loss3/loss14 = 0.308827 (* 0.0909091 = 0.0280752 loss) | |
| I0510 16:24:14.843984 10926 solver.cpp:245] Train net output #139: loss3/loss15 = 0.19023 (* 0.0909091 = 0.0172937 loss) | |
| I0510 16:24:14.843998 10926 solver.cpp:245] Train net output #140: loss3/loss16 = 0.122725 (* 0.0909091 = 0.0111569 loss) | |
| I0510 16:24:14.844012 10926 solver.cpp:245] Train net output #141: loss3/loss17 = 0.029164 (* 0.0909091 = 0.00265127 loss) | |
| I0510 16:24:14.844027 10926 solver.cpp:245] Train net output #142: loss3/loss18 = 0.0221107 (* 0.0909091 = 0.00201007 loss) | |
| I0510 16:24:14.844041 10926 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00630346 (* 0.0909091 = 0.000573041 loss) | |
| I0510 16:24:14.844055 10926 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00601129 (* 0.0909091 = 0.000546481 loss) | |
| I0510 16:24:14.844070 10926 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00325545 (* 0.0909091 = 0.00029595 loss) | |
| I0510 16:24:14.844084 10926 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00228187 (* 0.0909091 = 0.000207443 loss) | |
| I0510 16:24:14.844096 10926 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0510 16:24:14.844118 10926 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0510 16:24:14.844132 10926 solver.cpp:245] Train net output #149: total_confidence = 1.70201e-05 | |
| I0510 16:24:14.844144 10926 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.000268727 | |
| I0510 16:24:14.844157 10926 sgd_solver.cpp:106] Iteration 11500, lr = 0.001 | |
| I0510 16:24:36.112043 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 69.6994 > 30) by scale factor 0.43042 | |
| I0510 16:25:19.532807 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 31.874 > 30) by scale factor 0.941206 | |
| I0510 16:26:42.000691 10926 solver.cpp:229] Iteration 12000, loss = 10.1687 | |
| I0510 16:26:42.000823 10926 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0588235 | |
| I0510 16:26:42.000845 10926 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.25 | |
| I0510 16:26:42.000859 10926 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0510 16:26:42.000875 10926 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0510 16:26:42.000890 10926 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0 | |
| I0510 16:26:42.000902 10926 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.125 | |
| I0510 16:26:42.000916 10926 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.25 | |
| I0510 16:26:42.000931 10926 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.5 | |
| I0510 16:26:42.000942 10926 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.875 | |
| I0510 16:26:42.000955 10926 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0510 16:26:42.000968 10926 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0510 16:26:42.000982 10926 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0510 16:26:42.000999 10926 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0510 16:26:42.001013 10926 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0510 16:26:42.001026 10926 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0510 16:26:42.001039 10926 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0510 16:26:42.001050 10926 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0510 16:26:42.001062 10926 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0510 16:26:42.001075 10926 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0510 16:26:42.001086 10926 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0510 16:26:42.001098 10926 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0510 16:26:42.001111 10926 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0510 16:26:42.001142 10926 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0510 16:26:42.001158 10926 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.727273 | |
| I0510 16:26:42.001171 10926 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.196078 | |
| I0510 16:26:42.001188 10926 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.41728 (* 0.3 = 1.02518 loss) | |
| I0510 16:26:42.001204 10926 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.05673 (* 0.3 = 0.317018 loss) | |
| I0510 16:26:42.001219 10926 solver.cpp:245] Train net output #27: loss1/loss01 = 3.53676 (* 0.0272727 = 0.0964571 loss) | |
| I0510 16:26:42.001233 10926 solver.cpp:245] Train net output #28: loss1/loss02 = 3.47149 (* 0.0272727 = 0.0946771 loss) | |
| I0510 16:26:42.001248 10926 solver.cpp:245] Train net output #29: loss1/loss03 = 3.79591 (* 0.0272727 = 0.103525 loss) | |
| I0510 16:26:42.001263 10926 solver.cpp:245] Train net output #30: loss1/loss04 = 3.88131 (* 0.0272727 = 0.105854 loss) | |
| I0510 16:26:42.001277 10926 solver.cpp:245] Train net output #31: loss1/loss05 = 3.70254 (* 0.0272727 = 0.100978 loss) | |
| I0510 16:26:42.001292 10926 solver.cpp:245] Train net output #32: loss1/loss06 = 3.01672 (* 0.0272727 = 0.0822742 loss) | |
| I0510 16:26:42.001313 10926 solver.cpp:245] Train net output #33: loss1/loss07 = 2.94203 (* 0.0272727 = 0.0802371 loss) | |
| I0510 16:26:42.001327 10926 solver.cpp:245] Train net output #34: loss1/loss08 = 0.917379 (* 0.0272727 = 0.0250194 loss) | |
| I0510 16:26:42.001343 10926 solver.cpp:245] Train net output #35: loss1/loss09 = 0.0839761 (* 0.0272727 = 0.00229026 loss) | |
| I0510 16:26:42.001358 10926 solver.cpp:245] Train net output #36: loss1/loss10 = 0.0723426 (* 0.0272727 = 0.00197298 loss) | |
| I0510 16:26:42.001373 10926 solver.cpp:245] Train net output #37: loss1/loss11 = 0.0440893 (* 0.0272727 = 0.00120244 loss) | |
| I0510 16:26:42.001387 10926 solver.cpp:245] Train net output #38: loss1/loss12 = 0.0432924 (* 0.0272727 = 0.0011807 loss) | |
| I0510 16:26:42.001402 10926 solver.cpp:245] Train net output #39: loss1/loss13 = 0.0277258 (* 0.0272727 = 0.000756159 loss) | |
| I0510 16:26:42.001438 10926 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0291546 (* 0.0272727 = 0.000795124 loss) | |
| I0510 16:26:42.001454 10926 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0218412 (* 0.0272727 = 0.000595668 loss) | |
| I0510 16:26:42.001468 10926 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0200156 (* 0.0272727 = 0.000545879 loss) | |
| I0510 16:26:42.001484 10926 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0149774 (* 0.0272727 = 0.000408475 loss) | |
| I0510 16:26:42.001502 10926 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0209196 (* 0.0272727 = 0.000570536 loss) | |
| I0510 16:26:42.001516 10926 solver.cpp:245] Train net output #45: loss1/loss19 = 0.00990908 (* 0.0272727 = 0.000270248 loss) | |
| I0510 16:26:42.001531 10926 solver.cpp:245] Train net output #46: loss1/loss20 = 0.010353 (* 0.0272727 = 0.000282354 loss) | |
| I0510 16:26:42.001545 10926 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00872267 (* 0.0272727 = 0.000237891 loss) | |
| I0510 16:26:42.001560 10926 solver.cpp:245] Train net output #48: loss1/loss22 = 0.00468217 (* 0.0272727 = 0.000127696 loss) | |
| I0510 16:26:42.001574 10926 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0196078 | |
| I0510 16:26:42.001585 10926 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0510 16:26:42.001597 10926 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0.125 | |
| I0510 16:26:42.001610 10926 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0510 16:26:42.001621 10926 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.125 | |
| I0510 16:26:42.001633 10926 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0 | |
| I0510 16:26:42.001647 10926 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.25 | |
| I0510 16:26:42.001670 10926 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.5 | |
| I0510 16:26:42.001693 10926 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.875 | |
| I0510 16:26:42.001714 10926 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0510 16:26:42.001729 10926 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0510 16:26:42.001741 10926 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0510 16:26:42.001754 10926 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0510 16:26:42.001765 10926 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0510 16:26:42.001777 10926 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0510 16:26:42.001790 10926 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0510 16:26:42.001801 10926 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0510 16:26:42.001813 10926 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0510 16:26:42.001826 10926 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0510 16:26:42.001842 10926 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0510 16:26:42.001854 10926 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0510 16:26:42.001866 10926 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0510 16:26:42.001878 10926 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0510 16:26:42.001890 10926 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.715909 | |
| I0510 16:26:42.001904 10926 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.137255 | |
| I0510 16:26:42.001917 10926 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.41178 (* 0.3 = 1.02353 loss) | |
| I0510 16:26:42.001935 10926 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.02573 (* 0.3 = 0.307719 loss) | |
| I0510 16:26:42.001950 10926 solver.cpp:245] Train net output #76: loss2/loss01 = 3.43937 (* 0.0272727 = 0.0938011 loss) | |
| I0510 16:26:42.001963 10926 solver.cpp:245] Train net output #77: loss2/loss02 = 3.22131 (* 0.0272727 = 0.0878538 loss) | |
| I0510 16:26:42.001991 10926 solver.cpp:245] Train net output #78: loss2/loss03 = 3.1521 (* 0.0272727 = 0.0859664 loss) | |
| I0510 16:26:42.002007 10926 solver.cpp:245] Train net output #79: loss2/loss04 = 3.76162 (* 0.0272727 = 0.10259 loss) | |
| I0510 16:26:42.002022 10926 solver.cpp:245] Train net output #80: loss2/loss05 = 3.83694 (* 0.0272727 = 0.104644 loss) | |
| I0510 16:26:42.002035 10926 solver.cpp:245] Train net output #81: loss2/loss06 = 2.90377 (* 0.0272727 = 0.0791936 loss) | |
| I0510 16:26:42.002049 10926 solver.cpp:245] Train net output #82: loss2/loss07 = 2.68723 (* 0.0272727 = 0.0732881 loss) | |
| I0510 16:26:42.002063 10926 solver.cpp:245] Train net output #83: loss2/loss08 = 0.708419 (* 0.0272727 = 0.0193205 loss) | |
| I0510 16:26:42.002079 10926 solver.cpp:245] Train net output #84: loss2/loss09 = 0.0433044 (* 0.0272727 = 0.00118103 loss) | |
| I0510 16:26:42.002094 10926 solver.cpp:245] Train net output #85: loss2/loss10 = 0.0260731 (* 0.0272727 = 0.000711084 loss) | |
| I0510 16:26:42.002104 10926 solver.cpp:245] Train net output #86: loss2/loss11 = 0.0259263 (* 0.0272727 = 0.00070708 loss) | |
| I0510 16:26:42.002113 10926 solver.cpp:245] Train net output #87: loss2/loss12 = 0.0198567 (* 0.0272727 = 0.000541547 loss) | |
| I0510 16:26:42.002128 10926 solver.cpp:245] Train net output #88: loss2/loss13 = 0.0151725 (* 0.0272727 = 0.000413797 loss) | |
| I0510 16:26:42.002143 10926 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0188134 (* 0.0272727 = 0.000513094 loss) | |
| I0510 16:26:42.002157 10926 solver.cpp:245] Train net output #90: loss2/loss15 = 0.00999324 (* 0.0272727 = 0.000272543 loss) | |
| I0510 16:26:42.002172 10926 solver.cpp:245] Train net output #91: loss2/loss16 = 0.00772723 (* 0.0272727 = 0.000210743 loss) | |
| I0510 16:26:42.002187 10926 solver.cpp:245] Train net output #92: loss2/loss17 = 0.00825482 (* 0.0272727 = 0.000225131 loss) | |
| I0510 16:26:42.002208 10926 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0043138 (* 0.0272727 = 0.000117649 loss) | |
| I0510 16:26:42.002224 10926 solver.cpp:245] Train net output #94: loss2/loss19 = 0.00388719 (* 0.0272727 = 0.000106014 loss) | |
| I0510 16:26:42.002238 10926 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00369015 (* 0.0272727 = 0.000100641 loss) | |
| I0510 16:26:42.002254 10926 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00263287 (* 0.0272727 = 7.18056e-05 loss) | |
| I0510 16:26:42.002267 10926 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00384695 (* 0.0272727 = 0.000104917 loss) | |
| I0510 16:26:42.002280 10926 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0784314 | |
| I0510 16:26:42.002293 10926 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0510 16:26:42.002305 10926 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.125 | |
| I0510 16:26:42.002316 10926 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0510 16:26:42.002328 10926 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0 | |
| I0510 16:26:42.002341 10926 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0 | |
| I0510 16:26:42.002352 10926 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.25 | |
| I0510 16:26:42.002364 10926 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.5 | |
| I0510 16:26:42.002375 10926 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.875 | |
| I0510 16:26:42.002388 10926 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0510 16:26:42.002399 10926 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0510 16:26:42.002410 10926 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0510 16:26:42.002423 10926 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0510 16:26:42.002434 10926 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0510 16:26:42.002446 10926 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0510 16:26:42.002457 10926 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0510 16:26:42.002470 10926 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0510 16:26:42.002491 10926 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0510 16:26:42.002504 10926 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0510 16:26:42.002516 10926 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0510 16:26:42.002528 10926 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0510 16:26:42.002540 10926 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0510 16:26:42.002552 10926 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0510 16:26:42.002563 10926 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.732955 | |
| I0510 16:26:42.002575 10926 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.196078 | |
| I0510 16:26:42.002589 10926 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.11261 (* 1 = 3.11261 loss) | |
| I0510 16:26:42.002604 10926 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 0.953275 (* 1 = 0.953275 loss) | |
| I0510 16:26:42.002619 10926 solver.cpp:245] Train net output #125: loss3/loss01 = 3.02216 (* 0.0909091 = 0.274742 loss) | |
| I0510 16:26:42.002632 10926 solver.cpp:245] Train net output #126: loss3/loss02 = 2.93988 (* 0.0909091 = 0.267262 loss) | |
| I0510 16:26:42.002646 10926 solver.cpp:245] Train net output #127: loss3/loss03 = 3.17349 (* 0.0909091 = 0.288499 loss) | |
| I0510 16:26:42.002660 10926 solver.cpp:245] Train net output #128: loss3/loss04 = 3.18859 (* 0.0909091 = 0.289871 loss) | |
| I0510 16:26:42.002676 10926 solver.cpp:245] Train net output #129: loss3/loss05 = 3.37051 (* 0.0909091 = 0.30641 loss) | |
| I0510 16:26:42.002689 10926 solver.cpp:245] Train net output #130: loss3/loss06 = 2.54288 (* 0.0909091 = 0.231171 loss) | |
| I0510 16:26:42.002703 10926 solver.cpp:245] Train net output #131: loss3/loss07 = 2.16865 (* 0.0909091 = 0.19715 loss) | |
| I0510 16:26:42.002717 10926 solver.cpp:245] Train net output #132: loss3/loss08 = 0.629175 (* 0.0909091 = 0.0571977 loss) | |
| I0510 16:26:42.002732 10926 solver.cpp:245] Train net output #133: loss3/loss09 = 0.0345658 (* 0.0909091 = 0.00314234 loss) | |
| I0510 16:26:42.002746 10926 solver.cpp:245] Train net output #134: loss3/loss10 = 0.0257168 (* 0.0909091 = 0.00233789 loss) | |
| I0510 16:26:42.002760 10926 solver.cpp:245] Train net output #135: loss3/loss11 = 0.00899294 (* 0.0909091 = 0.00081754 loss) | |
| I0510 16:26:42.002774 10926 solver.cpp:245] Train net output #136: loss3/loss12 = 0.00692723 (* 0.0909091 = 0.000629748 loss) | |
| I0510 16:26:42.002789 10926 solver.cpp:245] Train net output #137: loss3/loss13 = 0.00631903 (* 0.0909091 = 0.000574457 loss) | |
| I0510 16:26:42.002804 10926 solver.cpp:245] Train net output #138: loss3/loss14 = 0.00651398 (* 0.0909091 = 0.00059218 loss) | |
| I0510 16:26:42.002817 10926 solver.cpp:245] Train net output #139: loss3/loss15 = 0.00361506 (* 0.0909091 = 0.000328642 loss) | |
| I0510 16:26:42.002831 10926 solver.cpp:245] Train net output #140: loss3/loss16 = 0.00325454 (* 0.0909091 = 0.000295867 loss) | |
| I0510 16:26:42.002846 10926 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00237569 (* 0.0909091 = 0.000215972 loss) | |
| I0510 16:26:42.002861 10926 solver.cpp:245] Train net output #142: loss3/loss18 = 0.0019862 (* 0.0909091 = 0.000180564 loss) | |
| I0510 16:26:42.002874 10926 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00123661 (* 0.0909091 = 0.000112419 loss) | |
| I0510 16:26:42.002892 10926 solver.cpp:245] Train net output #144: loss3/loss20 = 0.000716502 (* 0.0909091 = 6.51366e-05 loss) | |
| I0510 16:26:42.002908 10926 solver.cpp:245] Train net output #145: loss3/loss21 = 0.000821747 (* 0.0909091 = 7.47043e-05 loss) | |
| I0510 16:26:42.002923 10926 solver.cpp:245] Train net output #146: loss3/loss22 = 0.000613144 (* 0.0909091 = 5.57404e-05 loss) | |
| I0510 16:26:42.002936 10926 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0510 16:26:42.002948 10926 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0510 16:26:42.002964 10926 solver.cpp:245] Train net output #149: total_confidence = 1.21973e-06 | |
| I0510 16:26:42.003003 10926 solver.cpp:245] Train net output #150: total_confidence_not_rec = 1.67496e-05 | |
| I0510 16:26:42.003021 10926 sgd_solver.cpp:106] Iteration 12000, lr = 0.001 | |
| I0510 16:26:42.695148 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 30.6558 > 30) by scale factor 0.978608 | |
| I0510 16:26:56.245451 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 30.6004 > 30) by scale factor 0.980379 | |
| I0510 16:27:07.167090 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 43.6858 > 30) by scale factor 0.686722 | |
| I0510 16:28:36.391999 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 41.9066 > 30) by scale factor 0.715878 | |
| I0510 16:29:09.657678 10926 solver.cpp:229] Iteration 12500, loss = 10.2553 | |
| I0510 16:29:09.657757 10926 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0392157 | |
| I0510 16:29:09.657778 10926 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.125 | |
| I0510 16:29:09.657791 10926 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.125 | |
| I0510 16:29:09.657804 10926 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0510 16:29:09.657816 10926 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.25 | |
| I0510 16:29:09.657829 10926 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.75 | |
| I0510 16:29:09.657842 10926 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.5 | |
| I0510 16:29:09.657855 10926 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.625 | |
| I0510 16:29:09.657868 10926 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.75 | |
| I0510 16:29:09.657881 10926 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.75 | |
| I0510 16:29:09.657893 10926 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.75 | |
| I0510 16:29:09.657907 10926 solver.cpp:245] Train net output #11: loss1/accuracy11 = 0.875 | |
| I0510 16:29:09.657927 10926 solver.cpp:245] Train net output #12: loss1/accuracy12 = 0.875 | |
| I0510 16:29:09.657954 10926 solver.cpp:245] Train net output #13: loss1/accuracy13 = 0.875 | |
| I0510 16:29:09.657970 10926 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0510 16:29:09.657982 10926 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0510 16:29:09.657994 10926 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0510 16:29:09.658006 10926 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0510 16:29:09.658018 10926 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0510 16:29:09.658030 10926 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0510 16:29:09.658042 10926 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0510 16:29:09.658054 10926 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0510 16:29:09.658068 10926 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0510 16:29:09.658082 10926 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.710227 | |
| I0510 16:29:09.658093 10926 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.254902 | |
| I0510 16:29:09.658110 10926 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.05138 (* 0.3 = 0.915413 loss) | |
| I0510 16:29:09.658125 10926 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.14318 (* 0.3 = 0.342954 loss) | |
| I0510 16:29:09.658140 10926 solver.cpp:245] Train net output #27: loss1/loss01 = 3.4845 (* 0.0272727 = 0.0950318 loss) | |
| I0510 16:29:09.658154 10926 solver.cpp:245] Train net output #28: loss1/loss02 = 2.99473 (* 0.0272727 = 0.0816746 loss) | |
| I0510 16:29:09.658179 10926 solver.cpp:245] Train net output #29: loss1/loss03 = 3.83517 (* 0.0272727 = 0.104596 loss) | |
| I0510 16:29:09.658202 10926 solver.cpp:245] Train net output #30: loss1/loss04 = 3.15395 (* 0.0272727 = 0.0860168 loss) | |
| I0510 16:29:09.658217 10926 solver.cpp:245] Train net output #31: loss1/loss05 = 1.74047 (* 0.0272727 = 0.0474673 loss) | |
| I0510 16:29:09.658232 10926 solver.cpp:245] Train net output #32: loss1/loss06 = 1.87058 (* 0.0272727 = 0.0510157 loss) | |
| I0510 16:29:09.658246 10926 solver.cpp:245] Train net output #33: loss1/loss07 = 1.6615 (* 0.0272727 = 0.0453137 loss) | |
| I0510 16:29:09.658260 10926 solver.cpp:245] Train net output #34: loss1/loss08 = 0.912459 (* 0.0272727 = 0.0248853 loss) | |
| I0510 16:29:09.658275 10926 solver.cpp:245] Train net output #35: loss1/loss09 = 1.4229 (* 0.0272727 = 0.0388063 loss) | |
| I0510 16:29:09.658289 10926 solver.cpp:245] Train net output #36: loss1/loss10 = 0.891099 (* 0.0272727 = 0.0243027 loss) | |
| I0510 16:29:09.658304 10926 solver.cpp:245] Train net output #37: loss1/loss11 = 0.725622 (* 0.0272727 = 0.0197897 loss) | |
| I0510 16:29:09.658318 10926 solver.cpp:245] Train net output #38: loss1/loss12 = 0.563064 (* 0.0272727 = 0.0153563 loss) | |
| I0510 16:29:09.658334 10926 solver.cpp:245] Train net output #39: loss1/loss13 = 0.890479 (* 0.0272727 = 0.0242858 loss) | |
| I0510 16:29:09.658366 10926 solver.cpp:245] Train net output #40: loss1/loss14 = 0.132146 (* 0.0272727 = 0.00360397 loss) | |
| I0510 16:29:09.658382 10926 solver.cpp:245] Train net output #41: loss1/loss15 = 0.137137 (* 0.0272727 = 0.0037401 loss) | |
| I0510 16:29:09.658397 10926 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0445976 (* 0.0272727 = 0.0012163 loss) | |
| I0510 16:29:09.658412 10926 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0312428 (* 0.0272727 = 0.000852078 loss) | |
| I0510 16:29:09.658427 10926 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0139412 (* 0.0272727 = 0.000380215 loss) | |
| I0510 16:29:09.658442 10926 solver.cpp:245] Train net output #45: loss1/loss19 = 0.00871713 (* 0.0272727 = 0.00023774 loss) | |
| I0510 16:29:09.658457 10926 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0154528 (* 0.0272727 = 0.000421441 loss) | |
| I0510 16:29:09.658471 10926 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00545944 (* 0.0272727 = 0.000148894 loss) | |
| I0510 16:29:09.658485 10926 solver.cpp:245] Train net output #48: loss1/loss22 = 0.00780357 (* 0.0272727 = 0.000212825 loss) | |
| I0510 16:29:09.658499 10926 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0588235 | |
| I0510 16:29:09.658510 10926 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0510 16:29:09.658522 10926 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0.25 | |
| I0510 16:29:09.658535 10926 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0510 16:29:09.658546 10926 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.125 | |
| I0510 16:29:09.658558 10926 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.625 | |
| I0510 16:29:09.658571 10926 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.5 | |
| I0510 16:29:09.658582 10926 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.625 | |
| I0510 16:29:09.658594 10926 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.75 | |
| I0510 16:29:09.658607 10926 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.75 | |
| I0510 16:29:09.658618 10926 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.75 | |
| I0510 16:29:09.658630 10926 solver.cpp:245] Train net output #60: loss2/accuracy11 = 0.875 | |
| I0510 16:29:09.658641 10926 solver.cpp:245] Train net output #61: loss2/accuracy12 = 0.875 | |
| I0510 16:29:09.658654 10926 solver.cpp:245] Train net output #62: loss2/accuracy13 = 0.875 | |
| I0510 16:29:09.658665 10926 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0510 16:29:09.658677 10926 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0510 16:29:09.658689 10926 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0510 16:29:09.658700 10926 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0510 16:29:09.658715 10926 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0510 16:29:09.658727 10926 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0510 16:29:09.658740 10926 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0510 16:29:09.658751 10926 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0510 16:29:09.658761 10926 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0510 16:29:09.658773 10926 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.693182 | |
| I0510 16:29:09.658785 10926 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.196078 | |
| I0510 16:29:09.658799 10926 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.05444 (* 0.3 = 0.916332 loss) | |
| I0510 16:29:09.658813 10926 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.25251 (* 0.3 = 0.375754 loss) | |
| I0510 16:29:09.658828 10926 solver.cpp:245] Train net output #76: loss2/loss01 = 2.86258 (* 0.0272727 = 0.0780704 loss) | |
| I0510 16:29:09.658843 10926 solver.cpp:245] Train net output #77: loss2/loss02 = 3.32052 (* 0.0272727 = 0.0905598 loss) | |
| I0510 16:29:09.658871 10926 solver.cpp:245] Train net output #78: loss2/loss03 = 3.71964 (* 0.0272727 = 0.101445 loss) | |
| I0510 16:29:09.658888 10926 solver.cpp:245] Train net output #79: loss2/loss04 = 3.48845 (* 0.0272727 = 0.0951394 loss) | |
| I0510 16:29:09.658912 10926 solver.cpp:245] Train net output #80: loss2/loss05 = 1.95607 (* 0.0272727 = 0.0533475 loss) | |
| I0510 16:29:09.658934 10926 solver.cpp:245] Train net output #81: loss2/loss06 = 1.87358 (* 0.0272727 = 0.0510977 loss) | |
| I0510 16:29:09.658951 10926 solver.cpp:245] Train net output #82: loss2/loss07 = 1.63177 (* 0.0272727 = 0.0445028 loss) | |
| I0510 16:29:09.658964 10926 solver.cpp:245] Train net output #83: loss2/loss08 = 0.865461 (* 0.0272727 = 0.0236035 loss) | |
| I0510 16:29:09.658979 10926 solver.cpp:245] Train net output #84: loss2/loss09 = 1.11137 (* 0.0272727 = 0.0303101 loss) | |
| I0510 16:29:09.658993 10926 solver.cpp:245] Train net output #85: loss2/loss10 = 0.838645 (* 0.0272727 = 0.0228721 loss) | |
| I0510 16:29:09.659008 10926 solver.cpp:245] Train net output #86: loss2/loss11 = 0.671528 (* 0.0272727 = 0.0183144 loss) | |
| I0510 16:29:09.659023 10926 solver.cpp:245] Train net output #87: loss2/loss12 = 0.584542 (* 0.0272727 = 0.015942 loss) | |
| I0510 16:29:09.659037 10926 solver.cpp:245] Train net output #88: loss2/loss13 = 0.570357 (* 0.0272727 = 0.0155552 loss) | |
| I0510 16:29:09.659052 10926 solver.cpp:245] Train net output #89: loss2/loss14 = 0.078505 (* 0.0272727 = 0.00214104 loss) | |
| I0510 16:29:09.659066 10926 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0329583 (* 0.0272727 = 0.000898862 loss) | |
| I0510 16:29:09.659081 10926 solver.cpp:245] Train net output #91: loss2/loss16 = 0.03641 (* 0.0272727 = 0.000992999 loss) | |
| I0510 16:29:09.659096 10926 solver.cpp:245] Train net output #92: loss2/loss17 = 0.024457 (* 0.0272727 = 0.00066701 loss) | |
| I0510 16:29:09.659111 10926 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0198212 (* 0.0272727 = 0.000540577 loss) | |
| I0510 16:29:09.659128 10926 solver.cpp:245] Train net output #94: loss2/loss19 = 0.0176019 (* 0.0272727 = 0.000480051 loss) | |
| I0510 16:29:09.659143 10926 solver.cpp:245] Train net output #95: loss2/loss20 = 0.0125199 (* 0.0272727 = 0.000341453 loss) | |
| I0510 16:29:09.659157 10926 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00610976 (* 0.0272727 = 0.00016663 loss) | |
| I0510 16:29:09.659171 10926 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0121587 (* 0.0272727 = 0.000331601 loss) | |
| I0510 16:29:09.659184 10926 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0980392 | |
| I0510 16:29:09.659196 10926 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0510 16:29:09.659209 10926 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.125 | |
| I0510 16:29:09.659220 10926 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0510 16:29:09.659231 10926 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0510 16:29:09.659242 10926 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.5 | |
| I0510 16:29:09.659255 10926 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.5 | |
| I0510 16:29:09.659266 10926 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.625 | |
| I0510 16:29:09.659278 10926 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.75 | |
| I0510 16:29:09.659289 10926 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.75 | |
| I0510 16:29:09.659301 10926 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.75 | |
| I0510 16:29:09.659312 10926 solver.cpp:245] Train net output #109: loss3/accuracy11 = 0.875 | |
| I0510 16:29:09.659323 10926 solver.cpp:245] Train net output #110: loss3/accuracy12 = 0.875 | |
| I0510 16:29:09.659335 10926 solver.cpp:245] Train net output #111: loss3/accuracy13 = 0.875 | |
| I0510 16:29:09.659348 10926 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0510 16:29:09.659358 10926 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0510 16:29:09.659370 10926 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0510 16:29:09.659392 10926 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0510 16:29:09.659406 10926 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0510 16:29:09.659418 10926 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0510 16:29:09.659430 10926 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0510 16:29:09.659442 10926 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0510 16:29:09.659453 10926 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0510 16:29:09.659466 10926 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.732955 | |
| I0510 16:29:09.659477 10926 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.313726 | |
| I0510 16:29:09.659492 10926 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 2.93044 (* 1 = 2.93044 loss) | |
| I0510 16:29:09.659507 10926 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 0.999551 (* 1 = 0.999551 loss) | |
| I0510 16:29:09.659520 10926 solver.cpp:245] Train net output #125: loss3/loss01 = 3.07854 (* 0.0909091 = 0.279868 loss) | |
| I0510 16:29:09.659534 10926 solver.cpp:245] Train net output #126: loss3/loss02 = 2.79457 (* 0.0909091 = 0.254052 loss) | |
| I0510 16:29:09.659549 10926 solver.cpp:245] Train net output #127: loss3/loss03 = 3.08916 (* 0.0909091 = 0.280833 loss) | |
| I0510 16:29:09.659564 10926 solver.cpp:245] Train net output #128: loss3/loss04 = 2.92502 (* 0.0909091 = 0.265911 loss) | |
| I0510 16:29:09.659577 10926 solver.cpp:245] Train net output #129: loss3/loss05 = 1.65176 (* 0.0909091 = 0.15016 loss) | |
| I0510 16:29:09.659591 10926 solver.cpp:245] Train net output #130: loss3/loss06 = 1.88358 (* 0.0909091 = 0.171235 loss) | |
| I0510 16:29:09.659605 10926 solver.cpp:245] Train net output #131: loss3/loss07 = 1.2707 (* 0.0909091 = 0.115518 loss) | |
| I0510 16:29:09.659620 10926 solver.cpp:245] Train net output #132: loss3/loss08 = 0.804115 (* 0.0909091 = 0.0731014 loss) | |
| I0510 16:29:09.659633 10926 solver.cpp:245] Train net output #133: loss3/loss09 = 0.976672 (* 0.0909091 = 0.0887884 loss) | |
| I0510 16:29:09.659648 10926 solver.cpp:245] Train net output #134: loss3/loss10 = 0.784327 (* 0.0909091 = 0.0713024 loss) | |
| I0510 16:29:09.659662 10926 solver.cpp:245] Train net output #135: loss3/loss11 = 0.490523 (* 0.0909091 = 0.044593 loss) | |
| I0510 16:29:09.659677 10926 solver.cpp:245] Train net output #136: loss3/loss12 = 0.552409 (* 0.0909091 = 0.050219 loss) | |
| I0510 16:29:09.659690 10926 solver.cpp:245] Train net output #137: loss3/loss13 = 0.598527 (* 0.0909091 = 0.0544115 loss) | |
| I0510 16:29:09.659704 10926 solver.cpp:245] Train net output #138: loss3/loss14 = 0.125391 (* 0.0909091 = 0.0113992 loss) | |
| I0510 16:29:09.659719 10926 solver.cpp:245] Train net output #139: loss3/loss15 = 0.0616665 (* 0.0909091 = 0.00560604 loss) | |
| I0510 16:29:09.659734 10926 solver.cpp:245] Train net output #140: loss3/loss16 = 0.0311935 (* 0.0909091 = 0.00283577 loss) | |
| I0510 16:29:09.659747 10926 solver.cpp:245] Train net output #141: loss3/loss17 = 0.0195555 (* 0.0909091 = 0.00177777 loss) | |
| I0510 16:29:09.659765 10926 solver.cpp:245] Train net output #142: loss3/loss18 = 0.0091099 (* 0.0909091 = 0.000828173 loss) | |
| I0510 16:29:09.659780 10926 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00697105 (* 0.0909091 = 0.000633731 loss) | |
| I0510 16:29:09.659795 10926 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00442702 (* 0.0909091 = 0.000402457 loss) | |
| I0510 16:29:09.659808 10926 solver.cpp:245] Train net output #145: loss3/loss21 = 0.0021985 (* 0.0909091 = 0.000199864 loss) | |
| I0510 16:29:09.659823 10926 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00271339 (* 0.0909091 = 0.000246671 loss) | |
| I0510 16:29:09.659835 10926 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0510 16:29:09.659847 10926 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0510 16:29:09.659858 10926 solver.cpp:245] Train net output #149: total_confidence = 7.21395e-05 | |
| I0510 16:29:09.659881 10926 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.000516805 | |
| I0510 16:29:09.659896 10926 sgd_solver.cpp:106] Iteration 12500, lr = 0.001 | |
| I0510 16:29:22.165419 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 37.3032 > 30) by scale factor 0.804221 | |
| I0510 16:29:31.034893 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 31.4994 > 30) by scale factor 0.952398 | |
| I0510 16:29:31.921306 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 53.2079 > 30) by scale factor 0.563826 | |
| I0510 16:30:05.014255 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 36.9187 > 30) by scale factor 0.812595 | |
| I0510 16:30:18.609347 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 36.3412 > 30) by scale factor 0.825509 | |
| I0510 16:30:23.033277 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 30.2153 > 30) by scale factor 0.992874 | |
| I0510 16:31:01.718297 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 30.4312 > 30) by scale factor 0.985829 | |
| I0510 16:31:09.988821 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 30.92 > 30) by scale factor 0.970245 | |
| I0510 16:31:37.332890 10926 solver.cpp:229] Iteration 13000, loss = 10.2321 | |
| I0510 16:31:37.333029 10926 solver.cpp:245] Train net output #0: loss1/accuracy = 0.08 | |
| I0510 16:31:37.333050 10926 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.125 | |
| I0510 16:31:37.333065 10926 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0510 16:31:37.333078 10926 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.125 | |
| I0510 16:31:37.333097 10926 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.125 | |
| I0510 16:31:37.333111 10926 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.25 | |
| I0510 16:31:37.333123 10926 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.5 | |
| I0510 16:31:37.333137 10926 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.75 | |
| I0510 16:31:37.333170 10926 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.875 | |
| I0510 16:31:37.333184 10926 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.875 | |
| I0510 16:31:37.333197 10926 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.875 | |
| I0510 16:31:37.333210 10926 solver.cpp:245] Train net output #11: loss1/accuracy11 = 0.875 | |
| I0510 16:31:37.333232 10926 solver.cpp:245] Train net output #12: loss1/accuracy12 = 0.875 | |
| I0510 16:31:37.333245 10926 solver.cpp:245] Train net output #13: loss1/accuracy13 = 0.875 | |
| I0510 16:31:37.333257 10926 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0510 16:31:37.333269 10926 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0510 16:31:37.333281 10926 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0510 16:31:37.333293 10926 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0510 16:31:37.333305 10926 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0510 16:31:37.333317 10926 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0510 16:31:37.333328 10926 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0510 16:31:37.333339 10926 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0510 16:31:37.333353 10926 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0510 16:31:37.333364 10926 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.732955 | |
| I0510 16:31:37.333376 10926 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.24 | |
| I0510 16:31:37.333392 10926 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.20568 (* 0.3 = 0.961705 loss) | |
| I0510 16:31:37.333407 10926 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.05151 (* 0.3 = 0.315452 loss) | |
| I0510 16:31:37.333422 10926 solver.cpp:245] Train net output #27: loss1/loss01 = 3.51585 (* 0.0272727 = 0.0958867 loss) | |
| I0510 16:31:37.333436 10926 solver.cpp:245] Train net output #28: loss1/loss02 = 3.51615 (* 0.0272727 = 0.0958951 loss) | |
| I0510 16:31:37.333451 10926 solver.cpp:245] Train net output #29: loss1/loss03 = 3.11332 (* 0.0272727 = 0.0849087 loss) | |
| I0510 16:31:37.333466 10926 solver.cpp:245] Train net output #30: loss1/loss04 = 3.51172 (* 0.0272727 = 0.0957742 loss) | |
| I0510 16:31:37.333479 10926 solver.cpp:245] Train net output #31: loss1/loss05 = 2.9551 (* 0.0272727 = 0.0805936 loss) | |
| I0510 16:31:37.333494 10926 solver.cpp:245] Train net output #32: loss1/loss06 = 2.41916 (* 0.0272727 = 0.0659771 loss) | |
| I0510 16:31:37.333508 10926 solver.cpp:245] Train net output #33: loss1/loss07 = 1.4012 (* 0.0272727 = 0.0382145 loss) | |
| I0510 16:31:37.333523 10926 solver.cpp:245] Train net output #34: loss1/loss08 = 0.552252 (* 0.0272727 = 0.0150614 loss) | |
| I0510 16:31:37.333536 10926 solver.cpp:245] Train net output #35: loss1/loss09 = 0.461063 (* 0.0272727 = 0.0125744 loss) | |
| I0510 16:31:37.333551 10926 solver.cpp:245] Train net output #36: loss1/loss10 = 0.970798 (* 0.0272727 = 0.0264763 loss) | |
| I0510 16:31:37.333565 10926 solver.cpp:245] Train net output #37: loss1/loss11 = 0.601409 (* 0.0272727 = 0.0164021 loss) | |
| I0510 16:31:37.333580 10926 solver.cpp:245] Train net output #38: loss1/loss12 = 0.652265 (* 0.0272727 = 0.0177891 loss) | |
| I0510 16:31:37.333595 10926 solver.cpp:245] Train net output #39: loss1/loss13 = 0.701444 (* 0.0272727 = 0.0191303 loss) | |
| I0510 16:31:37.333766 10926 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0335508 (* 0.0272727 = 0.000915023 loss) | |
| I0510 16:31:37.333789 10926 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0210476 (* 0.0272727 = 0.000574026 loss) | |
| I0510 16:31:37.333806 10926 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0202086 (* 0.0272727 = 0.000551145 loss) | |
| I0510 16:31:37.333820 10926 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0147309 (* 0.0272727 = 0.000401751 loss) | |
| I0510 16:31:37.333835 10926 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0213879 (* 0.0272727 = 0.000583306 loss) | |
| I0510 16:31:37.333850 10926 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0130521 (* 0.0272727 = 0.000355965 loss) | |
| I0510 16:31:37.333865 10926 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0113298 (* 0.0272727 = 0.000308994 loss) | |
| I0510 16:31:37.333884 10926 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00975279 (* 0.0272727 = 0.000265985 loss) | |
| I0510 16:31:37.333899 10926 solver.cpp:245] Train net output #48: loss1/loss22 = 0.00774756 (* 0.0272727 = 0.000211297 loss) | |
| I0510 16:31:37.333912 10926 solver.cpp:245] Train net output #49: loss2/accuracy = 0.12 | |
| I0510 16:31:37.333925 10926 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.125 | |
| I0510 16:31:37.333938 10926 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0.125 | |
| I0510 16:31:37.333950 10926 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0510 16:31:37.333963 10926 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0 | |
| I0510 16:31:37.333976 10926 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.125 | |
| I0510 16:31:37.333987 10926 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.5 | |
| I0510 16:31:37.334000 10926 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.75 | |
| I0510 16:31:37.334012 10926 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.875 | |
| I0510 16:31:37.334025 10926 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.875 | |
| I0510 16:31:37.334038 10926 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.875 | |
| I0510 16:31:37.334049 10926 solver.cpp:245] Train net output #60: loss2/accuracy11 = 0.875 | |
| I0510 16:31:37.334062 10926 solver.cpp:245] Train net output #61: loss2/accuracy12 = 0.875 | |
| I0510 16:31:37.334074 10926 solver.cpp:245] Train net output #62: loss2/accuracy13 = 0.875 | |
| I0510 16:31:37.334086 10926 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0510 16:31:37.334098 10926 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0510 16:31:37.334110 10926 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0510 16:31:37.334122 10926 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0510 16:31:37.334134 10926 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0510 16:31:37.334146 10926 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0510 16:31:37.334157 10926 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0510 16:31:37.334169 10926 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0510 16:31:37.334182 10926 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0510 16:31:37.334193 10926 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.738636 | |
| I0510 16:31:37.334205 10926 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.24 | |
| I0510 16:31:37.334219 10926 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.29339 (* 0.3 = 0.988016 loss) | |
| I0510 16:31:37.334234 10926 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.09629 (* 0.3 = 0.328888 loss) | |
| I0510 16:31:37.334249 10926 solver.cpp:245] Train net output #76: loss2/loss01 = 3.46163 (* 0.0272727 = 0.0944081 loss) | |
| I0510 16:31:37.334264 10926 solver.cpp:245] Train net output #77: loss2/loss02 = 3.60205 (* 0.0272727 = 0.0982377 loss) | |
| I0510 16:31:37.334295 10926 solver.cpp:245] Train net output #78: loss2/loss03 = 3.17948 (* 0.0272727 = 0.0867132 loss) | |
| I0510 16:31:37.334326 10926 solver.cpp:245] Train net output #79: loss2/loss04 = 3.36413 (* 0.0272727 = 0.0917489 loss) | |
| I0510 16:31:37.334343 10926 solver.cpp:245] Train net output #80: loss2/loss05 = 3.59316 (* 0.0272727 = 0.0979954 loss) | |
| I0510 16:31:37.334363 10926 solver.cpp:245] Train net output #81: loss2/loss06 = 2.51491 (* 0.0272727 = 0.0685885 loss) | |
| I0510 16:31:37.334378 10926 solver.cpp:245] Train net output #82: loss2/loss07 = 1.11636 (* 0.0272727 = 0.0304461 loss) | |
| I0510 16:31:37.334393 10926 solver.cpp:245] Train net output #83: loss2/loss08 = 0.591282 (* 0.0272727 = 0.0161259 loss) | |
| I0510 16:31:37.334408 10926 solver.cpp:245] Train net output #84: loss2/loss09 = 0.318673 (* 0.0272727 = 0.00869108 loss) | |
| I0510 16:31:37.334422 10926 solver.cpp:245] Train net output #85: loss2/loss10 = 0.723906 (* 0.0272727 = 0.0197429 loss) | |
| I0510 16:31:37.334437 10926 solver.cpp:245] Train net output #86: loss2/loss11 = 0.46734 (* 0.0272727 = 0.0127456 loss) | |
| I0510 16:31:37.334451 10926 solver.cpp:245] Train net output #87: loss2/loss12 = 0.61822 (* 0.0272727 = 0.0168605 loss) | |
| I0510 16:31:37.334465 10926 solver.cpp:245] Train net output #88: loss2/loss13 = 0.85345 (* 0.0272727 = 0.0232759 loss) | |
| I0510 16:31:37.334481 10926 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0377375 (* 0.0272727 = 0.0010292 loss) | |
| I0510 16:31:37.334496 10926 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0310514 (* 0.0272727 = 0.000846856 loss) | |
| I0510 16:31:37.334509 10926 solver.cpp:245] Train net output #91: loss2/loss16 = 0.00565832 (* 0.0272727 = 0.000154318 loss) | |
| I0510 16:31:37.334524 10926 solver.cpp:245] Train net output #92: loss2/loss17 = 0.00628428 (* 0.0272727 = 0.000171389 loss) | |
| I0510 16:31:37.334538 10926 solver.cpp:245] Train net output #93: loss2/loss18 = 0.00517235 (* 0.0272727 = 0.000141064 loss) | |
| I0510 16:31:37.334553 10926 solver.cpp:245] Train net output #94: loss2/loss19 = 0.00335738 (* 0.0272727 = 9.1565e-05 loss) | |
| I0510 16:31:37.334568 10926 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00222298 (* 0.0272727 = 6.06269e-05 loss) | |
| I0510 16:31:37.334583 10926 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00212771 (* 0.0272727 = 5.80283e-05 loss) | |
| I0510 16:31:37.334594 10926 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00173729 (* 0.0272727 = 4.73805e-05 loss) | |
| I0510 16:31:37.334610 10926 solver.cpp:245] Train net output #98: loss3/accuracy = 0.14 | |
| I0510 16:31:37.334624 10926 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0510 16:31:37.334638 10926 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0510 16:31:37.334650 10926 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0.125 | |
| I0510 16:31:37.334662 10926 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0 | |
| I0510 16:31:37.334674 10926 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.125 | |
| I0510 16:31:37.334686 10926 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.375 | |
| I0510 16:31:37.334698 10926 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.75 | |
| I0510 16:31:37.334710 10926 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.875 | |
| I0510 16:31:37.334723 10926 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0510 16:31:37.334735 10926 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.875 | |
| I0510 16:31:37.334749 10926 solver.cpp:245] Train net output #109: loss3/accuracy11 = 0.875 | |
| I0510 16:31:37.334763 10926 solver.cpp:245] Train net output #110: loss3/accuracy12 = 0.875 | |
| I0510 16:31:37.334774 10926 solver.cpp:245] Train net output #111: loss3/accuracy13 = 0.875 | |
| I0510 16:31:37.334786 10926 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0510 16:31:37.334799 10926 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0510 16:31:37.334810 10926 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0510 16:31:37.334833 10926 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0510 16:31:37.334846 10926 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0510 16:31:37.334858 10926 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0510 16:31:37.334870 10926 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0510 16:31:37.334882 10926 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0510 16:31:37.334892 10926 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0510 16:31:37.334904 10926 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.732955 | |
| I0510 16:31:37.334916 10926 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.32 | |
| I0510 16:31:37.334933 10926 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.08672 (* 1 = 3.08672 loss) | |
| I0510 16:31:37.334947 10926 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.07204 (* 1 = 1.07204 loss) | |
| I0510 16:31:37.334962 10926 solver.cpp:245] Train net output #125: loss3/loss01 = 3.51843 (* 0.0909091 = 0.319857 loss) | |
| I0510 16:31:37.334976 10926 solver.cpp:245] Train net output #126: loss3/loss02 = 3.2793 (* 0.0909091 = 0.298119 loss) | |
| I0510 16:31:37.334990 10926 solver.cpp:245] Train net output #127: loss3/loss03 = 3.04704 (* 0.0909091 = 0.277003 loss) | |
| I0510 16:31:37.335005 10926 solver.cpp:245] Train net output #128: loss3/loss04 = 3.09923 (* 0.0909091 = 0.281748 loss) | |
| I0510 16:31:37.335019 10926 solver.cpp:245] Train net output #129: loss3/loss05 = 2.75595 (* 0.0909091 = 0.250541 loss) | |
| I0510 16:31:37.335033 10926 solver.cpp:245] Train net output #130: loss3/loss06 = 2.25977 (* 0.0909091 = 0.205433 loss) | |
| I0510 16:31:37.335047 10926 solver.cpp:245] Train net output #131: loss3/loss07 = 1.27346 (* 0.0909091 = 0.115769 loss) | |
| I0510 16:31:37.335062 10926 solver.cpp:245] Train net output #132: loss3/loss08 = 0.626191 (* 0.0909091 = 0.0569265 loss) | |
| I0510 16:31:37.335075 10926 solver.cpp:245] Train net output #133: loss3/loss09 = 0.257645 (* 0.0909091 = 0.0234222 loss) | |
| I0510 16:31:37.335089 10926 solver.cpp:245] Train net output #134: loss3/loss10 = 0.921412 (* 0.0909091 = 0.0837647 loss) | |
| I0510 16:31:37.335104 10926 solver.cpp:245] Train net output #135: loss3/loss11 = 0.434385 (* 0.0909091 = 0.0394895 loss) | |
| I0510 16:31:37.335119 10926 solver.cpp:245] Train net output #136: loss3/loss12 = 0.395414 (* 0.0909091 = 0.0359467 loss) | |
| I0510 16:31:37.335132 10926 solver.cpp:245] Train net output #137: loss3/loss13 = 0.461704 (* 0.0909091 = 0.0419731 loss) | |
| I0510 16:31:37.335146 10926 solver.cpp:245] Train net output #138: loss3/loss14 = 0.0319229 (* 0.0909091 = 0.00290208 loss) | |
| I0510 16:31:37.335161 10926 solver.cpp:245] Train net output #139: loss3/loss15 = 0.0253609 (* 0.0909091 = 0.00230554 loss) | |
| I0510 16:31:37.335175 10926 solver.cpp:245] Train net output #140: loss3/loss16 = 0.00941607 (* 0.0909091 = 0.000856006 loss) | |
| I0510 16:31:37.335191 10926 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00711306 (* 0.0909091 = 0.000646641 loss) | |
| I0510 16:31:37.335204 10926 solver.cpp:245] Train net output #142: loss3/loss18 = 0.0034142 (* 0.0909091 = 0.000310381 loss) | |
| I0510 16:31:37.335219 10926 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00208539 (* 0.0909091 = 0.000189581 loss) | |
| I0510 16:31:37.335233 10926 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00155358 (* 0.0909091 = 0.000141234 loss) | |
| I0510 16:31:37.335248 10926 solver.cpp:245] Train net output #145: loss3/loss21 = 0.000957421 (* 0.0909091 = 8.70383e-05 loss) | |
| I0510 16:31:37.335263 10926 solver.cpp:245] Train net output #146: loss3/loss22 = 0.000695691 (* 0.0909091 = 6.32447e-05 loss) | |
| I0510 16:31:37.335275 10926 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0510 16:31:37.335286 10926 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0510 16:31:37.335299 10926 solver.cpp:245] Train net output #149: total_confidence = 1.64442e-07 | |
| I0510 16:31:37.335321 10926 solver.cpp:245] Train net output #150: total_confidence_not_rec = 2.58883e-06 | |
| I0510 16:31:37.335336 10926 sgd_solver.cpp:106] Iteration 13000, lr = 0.001 | |
| I0510 16:33:09.410311 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 39.1697 > 30) by scale factor 0.765897 | |
| I0510 16:33:48.380039 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 51.5176 > 30) by scale factor 0.582325 | |
| I0510 16:34:00.491111 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 39.2076 > 30) by scale factor 0.765159 | |
| I0510 16:34:04.839540 10926 solver.cpp:229] Iteration 13500, loss = 10.1877 | |
| I0510 16:34:04.839613 10926 solver.cpp:245] Train net output #0: loss1/accuracy = 0.097561 | |
| I0510 16:34:04.839633 10926 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.125 | |
| I0510 16:34:04.839648 10926 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.25 | |
| I0510 16:34:04.839661 10926 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.125 | |
| I0510 16:34:04.839674 10926 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.25 | |
| I0510 16:34:04.839687 10926 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.25 | |
| I0510 16:34:04.839704 10926 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.625 | |
| I0510 16:34:04.839716 10926 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.875 | |
| I0510 16:34:04.839730 10926 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.875 | |
| I0510 16:34:04.839742 10926 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0510 16:34:04.839754 10926 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0510 16:34:04.839767 10926 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0510 16:34:04.839781 10926 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0510 16:34:04.839792 10926 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0510 16:34:04.839804 10926 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0510 16:34:04.839817 10926 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0510 16:34:04.839828 10926 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0510 16:34:04.839840 10926 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0510 16:34:04.839853 10926 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0510 16:34:04.839864 10926 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0510 16:34:04.839876 10926 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0510 16:34:04.839889 10926 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0510 16:34:04.839900 10926 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0510 16:34:04.839913 10926 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.772727 | |
| I0510 16:34:04.839926 10926 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.195122 | |
| I0510 16:34:04.839942 10926 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.68112 (* 0.3 = 1.10434 loss) | |
| I0510 16:34:04.839957 10926 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 0.996246 (* 0.3 = 0.298874 loss) | |
| I0510 16:34:04.839973 10926 solver.cpp:245] Train net output #27: loss1/loss01 = 3.64731 (* 0.0272727 = 0.0994721 loss) | |
| I0510 16:34:04.839988 10926 solver.cpp:245] Train net output #28: loss1/loss02 = 3.49673 (* 0.0272727 = 0.0953654 loss) | |
| I0510 16:34:04.840003 10926 solver.cpp:245] Train net output #29: loss1/loss03 = 3.57184 (* 0.0272727 = 0.0974137 loss) | |
| I0510 16:34:04.840018 10926 solver.cpp:245] Train net output #30: loss1/loss04 = 3.20117 (* 0.0272727 = 0.0873048 loss) | |
| I0510 16:34:04.840039 10926 solver.cpp:245] Train net output #31: loss1/loss05 = 3.23967 (* 0.0272727 = 0.0883547 loss) | |
| I0510 16:34:04.840054 10926 solver.cpp:245] Train net output #32: loss1/loss06 = 1.76084 (* 0.0272727 = 0.0480228 loss) | |
| I0510 16:34:04.840068 10926 solver.cpp:245] Train net output #33: loss1/loss07 = 0.691941 (* 0.0272727 = 0.0188711 loss) | |
| I0510 16:34:04.840082 10926 solver.cpp:245] Train net output #34: loss1/loss08 = 0.731482 (* 0.0272727 = 0.0199495 loss) | |
| I0510 16:34:04.840103 10926 solver.cpp:245] Train net output #35: loss1/loss09 = 0.0250535 (* 0.0272727 = 0.000683277 loss) | |
| I0510 16:34:04.840118 10926 solver.cpp:245] Train net output #36: loss1/loss10 = 0.0169293 (* 0.0272727 = 0.000461708 loss) | |
| I0510 16:34:04.840132 10926 solver.cpp:245] Train net output #37: loss1/loss11 = 0.0135043 (* 0.0272727 = 0.0003683 loss) | |
| I0510 16:34:04.840186 10926 solver.cpp:245] Train net output #38: loss1/loss12 = 0.00978179 (* 0.0272727 = 0.000266776 loss) | |
| I0510 16:34:04.840203 10926 solver.cpp:245] Train net output #39: loss1/loss13 = 0.00542027 (* 0.0272727 = 0.000147826 loss) | |
| I0510 16:34:04.840216 10926 solver.cpp:245] Train net output #40: loss1/loss14 = 0.00476109 (* 0.0272727 = 0.000129848 loss) | |
| I0510 16:34:04.840240 10926 solver.cpp:245] Train net output #41: loss1/loss15 = 0.00773909 (* 0.0272727 = 0.000211066 loss) | |
| I0510 16:34:04.840255 10926 solver.cpp:245] Train net output #42: loss1/loss16 = 0.00484537 (* 0.0272727 = 0.000132146 loss) | |
| I0510 16:34:04.840270 10926 solver.cpp:245] Train net output #43: loss1/loss17 = 0.00167026 (* 0.0272727 = 4.55526e-05 loss) | |
| I0510 16:34:04.840283 10926 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0013421 (* 0.0272727 = 3.66027e-05 loss) | |
| I0510 16:34:04.840298 10926 solver.cpp:245] Train net output #45: loss1/loss19 = 0.00168069 (* 0.0272727 = 4.5837e-05 loss) | |
| I0510 16:34:04.840313 10926 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00170694 (* 0.0272727 = 4.65528e-05 loss) | |
| I0510 16:34:04.840327 10926 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00129272 (* 0.0272727 = 3.52559e-05 loss) | |
| I0510 16:34:04.840342 10926 solver.cpp:245] Train net output #48: loss1/loss22 = 0.00123756 (* 0.0272727 = 3.37516e-05 loss) | |
| I0510 16:34:04.840354 10926 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0243902 | |
| I0510 16:34:04.840368 10926 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0510 16:34:04.840380 10926 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0510 16:34:04.840391 10926 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0510 16:34:04.840404 10926 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.375 | |
| I0510 16:34:04.840415 10926 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.25 | |
| I0510 16:34:04.840427 10926 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.625 | |
| I0510 16:34:04.840443 10926 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.875 | |
| I0510 16:34:04.840456 10926 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.875 | |
| I0510 16:34:04.840468 10926 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0510 16:34:04.840481 10926 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0510 16:34:04.840493 10926 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0510 16:34:04.840504 10926 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0510 16:34:04.840517 10926 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0510 16:34:04.840528 10926 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0510 16:34:04.840540 10926 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0510 16:34:04.840551 10926 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0510 16:34:04.840564 10926 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0510 16:34:04.840574 10926 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0510 16:34:04.840586 10926 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0510 16:34:04.840598 10926 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0510 16:34:04.840610 10926 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0510 16:34:04.840621 10926 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0510 16:34:04.840632 10926 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.772727 | |
| I0510 16:34:04.840644 10926 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.097561 | |
| I0510 16:34:04.840659 10926 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.78182 (* 0.3 = 1.13455 loss) | |
| I0510 16:34:04.840674 10926 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 0.964441 (* 0.3 = 0.289332 loss) | |
| I0510 16:34:04.840699 10926 solver.cpp:245] Train net output #76: loss2/loss01 = 4.01062 (* 0.0272727 = 0.109381 loss) | |
| I0510 16:34:04.840716 10926 solver.cpp:245] Train net output #77: loss2/loss02 = 3.69495 (* 0.0272727 = 0.100771 loss) | |
| I0510 16:34:04.840731 10926 solver.cpp:245] Train net output #78: loss2/loss03 = 3.89422 (* 0.0272727 = 0.106206 loss) | |
| I0510 16:34:04.840744 10926 solver.cpp:245] Train net output #79: loss2/loss04 = 3.28527 (* 0.0272727 = 0.0895983 loss) | |
| I0510 16:34:04.840759 10926 solver.cpp:245] Train net output #80: loss2/loss05 = 3.4536 (* 0.0272727 = 0.0941891 loss) | |
| I0510 16:34:04.840775 10926 solver.cpp:245] Train net output #81: loss2/loss06 = 1.68258 (* 0.0272727 = 0.0458886 loss) | |
| I0510 16:34:04.840788 10926 solver.cpp:245] Train net output #82: loss2/loss07 = 0.819929 (* 0.0272727 = 0.0223617 loss) | |
| I0510 16:34:04.840802 10926 solver.cpp:245] Train net output #83: loss2/loss08 = 0.655826 (* 0.0272727 = 0.0178862 loss) | |
| I0510 16:34:04.840817 10926 solver.cpp:245] Train net output #84: loss2/loss09 = 0.0314962 (* 0.0272727 = 0.000858986 loss) | |
| I0510 16:34:04.840832 10926 solver.cpp:245] Train net output #85: loss2/loss10 = 0.0265046 (* 0.0272727 = 0.000722853 loss) | |
| I0510 16:34:04.840847 10926 solver.cpp:245] Train net output #86: loss2/loss11 = 0.0132478 (* 0.0272727 = 0.000361305 loss) | |
| I0510 16:34:04.840860 10926 solver.cpp:245] Train net output #87: loss2/loss12 = 0.0184896 (* 0.0272727 = 0.000504261 loss) | |
| I0510 16:34:04.840875 10926 solver.cpp:245] Train net output #88: loss2/loss13 = 0.010064 (* 0.0272727 = 0.000274472 loss) | |
| I0510 16:34:04.840889 10926 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0102029 (* 0.0272727 = 0.000278261 loss) | |
| I0510 16:34:04.840904 10926 solver.cpp:245] Train net output #90: loss2/loss15 = 0.00738705 (* 0.0272727 = 0.000201465 loss) | |
| I0510 16:34:04.840919 10926 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0176342 (* 0.0272727 = 0.000480933 loss) | |
| I0510 16:34:04.840932 10926 solver.cpp:245] Train net output #92: loss2/loss17 = 0.00616881 (* 0.0272727 = 0.00016824 loss) | |
| I0510 16:34:04.840947 10926 solver.cpp:245] Train net output #93: loss2/loss18 = 0.00277049 (* 0.0272727 = 7.55588e-05 loss) | |
| I0510 16:34:04.840962 10926 solver.cpp:245] Train net output #94: loss2/loss19 = 0.00194466 (* 0.0272727 = 5.30361e-05 loss) | |
| I0510 16:34:04.840976 10926 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00359711 (* 0.0272727 = 9.81031e-05 loss) | |
| I0510 16:34:04.840991 10926 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00266201 (* 0.0272727 = 7.26003e-05 loss) | |
| I0510 16:34:04.841004 10926 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00263487 (* 0.0272727 = 7.18601e-05 loss) | |
| I0510 16:34:04.841017 10926 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0243902 | |
| I0510 16:34:04.841029 10926 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0510 16:34:04.841042 10926 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.125 | |
| I0510 16:34:04.841053 10926 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0510 16:34:04.841065 10926 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.25 | |
| I0510 16:34:04.841076 10926 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.25 | |
| I0510 16:34:04.841089 10926 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.625 | |
| I0510 16:34:04.841100 10926 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.875 | |
| I0510 16:34:04.841111 10926 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.875 | |
| I0510 16:34:04.841136 10926 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0510 16:34:04.841150 10926 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0510 16:34:04.841166 10926 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0510 16:34:04.841177 10926 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0510 16:34:04.841189 10926 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0510 16:34:04.841212 10926 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0510 16:34:04.841225 10926 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0510 16:34:04.841243 10926 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0510 16:34:04.841253 10926 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0510 16:34:04.841265 10926 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0510 16:34:04.841277 10926 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0510 16:34:04.841289 10926 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0510 16:34:04.841300 10926 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0510 16:34:04.841312 10926 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0510 16:34:04.841325 10926 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.767045 | |
| I0510 16:34:04.841336 10926 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.146341 | |
| I0510 16:34:04.841351 10926 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.55605 (* 1 = 3.55605 loss) | |
| I0510 16:34:04.841364 10926 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 0.920372 (* 1 = 0.920372 loss) | |
| I0510 16:34:04.841379 10926 solver.cpp:245] Train net output #125: loss3/loss01 = 3.64283 (* 0.0909091 = 0.331166 loss) | |
| I0510 16:34:04.841394 10926 solver.cpp:245] Train net output #126: loss3/loss02 = 3.34893 (* 0.0909091 = 0.304448 loss) | |
| I0510 16:34:04.841408 10926 solver.cpp:245] Train net output #127: loss3/loss03 = 3.84765 (* 0.0909091 = 0.349786 loss) | |
| I0510 16:34:04.841423 10926 solver.cpp:245] Train net output #128: loss3/loss04 = 3.10234 (* 0.0909091 = 0.282031 loss) | |
| I0510 16:34:04.841436 10926 solver.cpp:245] Train net output #129: loss3/loss05 = 2.93766 (* 0.0909091 = 0.26706 loss) | |
| I0510 16:34:04.841451 10926 solver.cpp:245] Train net output #130: loss3/loss06 = 1.51041 (* 0.0909091 = 0.13731 loss) | |
| I0510 16:34:04.841465 10926 solver.cpp:245] Train net output #131: loss3/loss07 = 0.529389 (* 0.0909091 = 0.0481263 loss) | |
| I0510 16:34:04.841480 10926 solver.cpp:245] Train net output #132: loss3/loss08 = 0.708514 (* 0.0909091 = 0.0644103 loss) | |
| I0510 16:34:04.841497 10926 solver.cpp:245] Train net output #133: loss3/loss09 = 0.0314208 (* 0.0909091 = 0.00285643 loss) | |
| I0510 16:34:04.841513 10926 solver.cpp:245] Train net output #134: loss3/loss10 = 0.0129425 (* 0.0909091 = 0.00117659 loss) | |
| I0510 16:34:04.841527 10926 solver.cpp:245] Train net output #135: loss3/loss11 = 0.00672807 (* 0.0909091 = 0.000611642 loss) | |
| I0510 16:34:04.841542 10926 solver.cpp:245] Train net output #136: loss3/loss12 = 0.00567915 (* 0.0909091 = 0.000516287 loss) | |
| I0510 16:34:04.841557 10926 solver.cpp:245] Train net output #137: loss3/loss13 = 0.00466079 (* 0.0909091 = 0.000423708 loss) | |
| I0510 16:34:04.841570 10926 solver.cpp:245] Train net output #138: loss3/loss14 = 0.00337436 (* 0.0909091 = 0.00030676 loss) | |
| I0510 16:34:04.841585 10926 solver.cpp:245] Train net output #139: loss3/loss15 = 0.00275185 (* 0.0909091 = 0.000250168 loss) | |
| I0510 16:34:04.841599 10926 solver.cpp:245] Train net output #140: loss3/loss16 = 0.00344023 (* 0.0909091 = 0.000312748 loss) | |
| I0510 16:34:04.841614 10926 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00172976 (* 0.0909091 = 0.000157251 loss) | |
| I0510 16:34:04.841629 10926 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00152216 (* 0.0909091 = 0.000138378 loss) | |
| I0510 16:34:04.841642 10926 solver.cpp:245] Train net output #143: loss3/loss19 = 0.0015925 (* 0.0909091 = 0.000144772 loss) | |
| I0510 16:34:04.841656 10926 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00116166 (* 0.0909091 = 0.000105605 loss) | |
| I0510 16:34:04.841671 10926 solver.cpp:245] Train net output #145: loss3/loss21 = 0.000890482 (* 0.0909091 = 8.09529e-05 loss) | |
| I0510 16:34:04.841686 10926 solver.cpp:245] Train net output #146: loss3/loss22 = 0.000982929 (* 0.0909091 = 8.93572e-05 loss) | |
| I0510 16:34:04.841707 10926 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0510 16:34:04.841722 10926 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0510 16:34:04.841733 10926 solver.cpp:245] Train net output #149: total_confidence = 0.000139729 | |
| I0510 16:34:04.841745 10926 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.000913595 | |
| I0510 16:34:04.841759 10926 sgd_solver.cpp:106] Iteration 13500, lr = 0.001 | |
| I0510 16:34:23.553680 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 30.8263 > 30) by scale factor 0.973194 | |
| I0510 16:34:28.269696 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 30.2988 > 30) by scale factor 0.990137 | |
| I0510 16:35:51.130724 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 31.7567 > 30) by scale factor 0.944683 | |
| I0510 16:36:32.140278 10926 solver.cpp:229] Iteration 14000, loss = 10.0424 | |
| I0510 16:36:32.140401 10926 solver.cpp:245] Train net output #0: loss1/accuracy = 0.108696 | |
| I0510 16:36:32.140422 10926 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0510 16:36:32.140436 10926 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.125 | |
| I0510 16:36:32.140450 10926 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.125 | |
| I0510 16:36:32.140462 10926 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.25 | |
| I0510 16:36:32.140475 10926 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.625 | |
| I0510 16:36:32.140488 10926 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.625 | |
| I0510 16:36:32.140501 10926 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.625 | |
| I0510 16:36:32.140513 10926 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.75 | |
| I0510 16:36:32.140525 10926 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.875 | |
| I0510 16:36:32.140538 10926 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.875 | |
| I0510 16:36:32.140550 10926 solver.cpp:245] Train net output #11: loss1/accuracy11 = 0.875 | |
| I0510 16:36:32.140563 10926 solver.cpp:245] Train net output #12: loss1/accuracy12 = 0.875 | |
| I0510 16:36:32.140575 10926 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0510 16:36:32.140588 10926 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0510 16:36:32.140599 10926 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0510 16:36:32.140611 10926 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0510 16:36:32.140624 10926 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0510 16:36:32.140635 10926 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0510 16:36:32.140650 10926 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0510 16:36:32.140661 10926 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0510 16:36:32.140674 10926 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0510 16:36:32.140686 10926 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0510 16:36:32.140698 10926 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.755682 | |
| I0510 16:36:32.140710 10926 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.195652 | |
| I0510 16:36:32.140727 10926 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.04714 (* 0.3 = 0.914142 loss) | |
| I0510 16:36:32.140743 10926 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 0.912007 (* 0.3 = 0.273602 loss) | |
| I0510 16:36:32.140758 10926 solver.cpp:245] Train net output #27: loss1/loss01 = 3.2262 (* 0.0272727 = 0.0879874 loss) | |
| I0510 16:36:32.140771 10926 solver.cpp:245] Train net output #28: loss1/loss02 = 3.59089 (* 0.0272727 = 0.0979334 loss) | |
| I0510 16:36:32.140785 10926 solver.cpp:245] Train net output #29: loss1/loss03 = 3.37428 (* 0.0272727 = 0.0920258 loss) | |
| I0510 16:36:32.140799 10926 solver.cpp:245] Train net output #30: loss1/loss04 = 2.40485 (* 0.0272727 = 0.0655868 loss) | |
| I0510 16:36:32.140813 10926 solver.cpp:245] Train net output #31: loss1/loss05 = 1.52465 (* 0.0272727 = 0.0415814 loss) | |
| I0510 16:36:32.140828 10926 solver.cpp:245] Train net output #32: loss1/loss06 = 1.57859 (* 0.0272727 = 0.0430525 loss) | |
| I0510 16:36:32.140841 10926 solver.cpp:245] Train net output #33: loss1/loss07 = 1.69241 (* 0.0272727 = 0.0461568 loss) | |
| I0510 16:36:32.140856 10926 solver.cpp:245] Train net output #34: loss1/loss08 = 1.70558 (* 0.0272727 = 0.0465159 loss) | |
| I0510 16:36:32.140871 10926 solver.cpp:245] Train net output #35: loss1/loss09 = 0.390813 (* 0.0272727 = 0.0106585 loss) | |
| I0510 16:36:32.140884 10926 solver.cpp:245] Train net output #36: loss1/loss10 = 0.572568 (* 0.0272727 = 0.0156155 loss) | |
| I0510 16:36:32.140898 10926 solver.cpp:245] Train net output #37: loss1/loss11 = 0.529043 (* 0.0272727 = 0.0144284 loss) | |
| I0510 16:36:32.140913 10926 solver.cpp:245] Train net output #38: loss1/loss12 = 0.604349 (* 0.0272727 = 0.0164822 loss) | |
| I0510 16:36:32.140946 10926 solver.cpp:245] Train net output #39: loss1/loss13 = 0.0436796 (* 0.0272727 = 0.00119126 loss) | |
| I0510 16:36:32.140962 10926 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0150694 (* 0.0272727 = 0.000410983 loss) | |
| I0510 16:36:32.140977 10926 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0207455 (* 0.0272727 = 0.000565787 loss) | |
| I0510 16:36:32.140992 10926 solver.cpp:245] Train net output #42: loss1/loss16 = 0.00663515 (* 0.0272727 = 0.000180959 loss) | |
| I0510 16:36:32.141007 10926 solver.cpp:245] Train net output #43: loss1/loss17 = 0.00356457 (* 0.0272727 = 9.72154e-05 loss) | |
| I0510 16:36:32.141021 10926 solver.cpp:245] Train net output #44: loss1/loss18 = 0.00339561 (* 0.0272727 = 9.26075e-05 loss) | |
| I0510 16:36:32.141036 10926 solver.cpp:245] Train net output #45: loss1/loss19 = 0.00207 (* 0.0272727 = 5.64544e-05 loss) | |
| I0510 16:36:32.141049 10926 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00275847 (* 0.0272727 = 7.5231e-05 loss) | |
| I0510 16:36:32.141063 10926 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00165557 (* 0.0272727 = 4.5152e-05 loss) | |
| I0510 16:36:32.141078 10926 solver.cpp:245] Train net output #48: loss1/loss22 = 0.00149849 (* 0.0272727 = 4.08678e-05 loss) | |
| I0510 16:36:32.141090 10926 solver.cpp:245] Train net output #49: loss2/accuracy = 0.108696 | |
| I0510 16:36:32.141103 10926 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0510 16:36:32.141115 10926 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0.125 | |
| I0510 16:36:32.141142 10926 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0.125 | |
| I0510 16:36:32.141155 10926 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.25 | |
| I0510 16:36:32.141167 10926 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.625 | |
| I0510 16:36:32.141180 10926 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.5 | |
| I0510 16:36:32.141191 10926 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.625 | |
| I0510 16:36:32.141203 10926 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.75 | |
| I0510 16:36:32.141214 10926 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.875 | |
| I0510 16:36:32.141227 10926 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.875 | |
| I0510 16:36:32.141238 10926 solver.cpp:245] Train net output #60: loss2/accuracy11 = 0.875 | |
| I0510 16:36:32.141250 10926 solver.cpp:245] Train net output #61: loss2/accuracy12 = 0.875 | |
| I0510 16:36:32.141261 10926 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0510 16:36:32.141273 10926 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0510 16:36:32.141285 10926 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0510 16:36:32.141296 10926 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0510 16:36:32.141309 10926 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0510 16:36:32.141319 10926 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0510 16:36:32.141331 10926 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0510 16:36:32.141342 10926 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0510 16:36:32.141355 10926 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0510 16:36:32.141366 10926 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0510 16:36:32.141377 10926 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.744318 | |
| I0510 16:36:32.141389 10926 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.173913 | |
| I0510 16:36:32.141403 10926 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.03455 (* 0.3 = 0.910364 loss) | |
| I0510 16:36:32.141418 10926 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 0.986398 (* 0.3 = 0.295919 loss) | |
| I0510 16:36:32.141433 10926 solver.cpp:245] Train net output #76: loss2/loss01 = 3.11895 (* 0.0272727 = 0.0850623 loss) | |
| I0510 16:36:32.141443 10926 solver.cpp:245] Train net output #77: loss2/loss02 = 3.01313 (* 0.0272727 = 0.0821763 loss) | |
| I0510 16:36:32.141474 10926 solver.cpp:245] Train net output #78: loss2/loss03 = 3.37721 (* 0.0272727 = 0.0921058 loss) | |
| I0510 16:36:32.141490 10926 solver.cpp:245] Train net output #79: loss2/loss04 = 2.9878 (* 0.0272727 = 0.0814855 loss) | |
| I0510 16:36:32.141505 10926 solver.cpp:245] Train net output #80: loss2/loss05 = 1.82466 (* 0.0272727 = 0.0497634 loss) | |
| I0510 16:36:32.141520 10926 solver.cpp:245] Train net output #81: loss2/loss06 = 1.95896 (* 0.0272727 = 0.0534262 loss) | |
| I0510 16:36:32.141533 10926 solver.cpp:245] Train net output #82: loss2/loss07 = 1.54149 (* 0.0272727 = 0.0420406 loss) | |
| I0510 16:36:32.141547 10926 solver.cpp:245] Train net output #83: loss2/loss08 = 1.44565 (* 0.0272727 = 0.0394269 loss) | |
| I0510 16:36:32.141561 10926 solver.cpp:245] Train net output #84: loss2/loss09 = 0.390029 (* 0.0272727 = 0.0106371 loss) | |
| I0510 16:36:32.141576 10926 solver.cpp:245] Train net output #85: loss2/loss10 = 0.271324 (* 0.0272727 = 0.00739974 loss) | |
| I0510 16:36:32.141590 10926 solver.cpp:245] Train net output #86: loss2/loss11 = 0.512969 (* 0.0272727 = 0.0139901 loss) | |
| I0510 16:36:32.141604 10926 solver.cpp:245] Train net output #87: loss2/loss12 = 0.397108 (* 0.0272727 = 0.0108302 loss) | |
| I0510 16:36:32.141618 10926 solver.cpp:245] Train net output #88: loss2/loss13 = 0.0707213 (* 0.0272727 = 0.00192876 loss) | |
| I0510 16:36:32.141633 10926 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0458379 (* 0.0272727 = 0.00125012 loss) | |
| I0510 16:36:32.141647 10926 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0494263 (* 0.0272727 = 0.00134799 loss) | |
| I0510 16:36:32.141661 10926 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0105051 (* 0.0272727 = 0.000286502 loss) | |
| I0510 16:36:32.141675 10926 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0203641 (* 0.0272727 = 0.000555384 loss) | |
| I0510 16:36:32.141690 10926 solver.cpp:245] Train net output #93: loss2/loss18 = 0.00607877 (* 0.0272727 = 0.000165785 loss) | |
| I0510 16:36:32.141706 10926 solver.cpp:245] Train net output #94: loss2/loss19 = 0.00573569 (* 0.0272727 = 0.000156428 loss) | |
| I0510 16:36:32.141721 10926 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00330532 (* 0.0272727 = 9.0145e-05 loss) | |
| I0510 16:36:32.141736 10926 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00294791 (* 0.0272727 = 8.03977e-05 loss) | |
| I0510 16:36:32.141749 10926 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00183376 (* 0.0272727 = 5.00116e-05 loss) | |
| I0510 16:36:32.141762 10926 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0652174 | |
| I0510 16:36:32.141774 10926 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0510 16:36:32.141787 10926 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0510 16:36:32.141798 10926 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0510 16:36:32.141809 10926 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.375 | |
| I0510 16:36:32.141820 10926 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.75 | |
| I0510 16:36:32.141832 10926 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.5 | |
| I0510 16:36:32.141844 10926 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.625 | |
| I0510 16:36:32.141855 10926 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.75 | |
| I0510 16:36:32.141867 10926 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.875 | |
| I0510 16:36:32.141878 10926 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.875 | |
| I0510 16:36:32.141891 10926 solver.cpp:245] Train net output #109: loss3/accuracy11 = 0.875 | |
| I0510 16:36:32.141901 10926 solver.cpp:245] Train net output #110: loss3/accuracy12 = 0.875 | |
| I0510 16:36:32.141913 10926 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0510 16:36:32.141924 10926 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0510 16:36:32.141937 10926 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0510 16:36:32.141947 10926 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0510 16:36:32.141969 10926 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0510 16:36:32.141983 10926 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0510 16:36:32.141994 10926 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0510 16:36:32.142006 10926 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0510 16:36:32.142019 10926 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0510 16:36:32.142030 10926 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0510 16:36:32.142041 10926 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.744318 | |
| I0510 16:36:32.142053 10926 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.217391 | |
| I0510 16:36:32.142067 10926 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 2.9844 (* 1 = 2.9844 loss) | |
| I0510 16:36:32.142081 10926 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 0.864491 (* 1 = 0.864491 loss) | |
| I0510 16:36:32.142096 10926 solver.cpp:245] Train net output #125: loss3/loss01 = 3.01046 (* 0.0909091 = 0.273678 loss) | |
| I0510 16:36:32.142109 10926 solver.cpp:245] Train net output #126: loss3/loss02 = 2.97102 (* 0.0909091 = 0.270093 loss) | |
| I0510 16:36:32.142123 10926 solver.cpp:245] Train net output #127: loss3/loss03 = 2.95608 (* 0.0909091 = 0.268735 loss) | |
| I0510 16:36:32.142138 10926 solver.cpp:245] Train net output #128: loss3/loss04 = 2.50734 (* 0.0909091 = 0.22794 loss) | |
| I0510 16:36:32.142151 10926 solver.cpp:245] Train net output #129: loss3/loss05 = 1.18704 (* 0.0909091 = 0.107913 loss) | |
| I0510 16:36:32.142165 10926 solver.cpp:245] Train net output #130: loss3/loss06 = 1.51364 (* 0.0909091 = 0.137603 loss) | |
| I0510 16:36:32.142179 10926 solver.cpp:245] Train net output #131: loss3/loss07 = 1.35909 (* 0.0909091 = 0.123553 loss) | |
| I0510 16:36:32.142194 10926 solver.cpp:245] Train net output #132: loss3/loss08 = 1.01427 (* 0.0909091 = 0.0922063 loss) | |
| I0510 16:36:32.142207 10926 solver.cpp:245] Train net output #133: loss3/loss09 = 0.340712 (* 0.0909091 = 0.0309738 loss) | |
| I0510 16:36:32.142221 10926 solver.cpp:245] Train net output #134: loss3/loss10 = 0.372195 (* 0.0909091 = 0.0338359 loss) | |
| I0510 16:36:32.142235 10926 solver.cpp:245] Train net output #135: loss3/loss11 = 0.581495 (* 0.0909091 = 0.0528632 loss) | |
| I0510 16:36:32.142249 10926 solver.cpp:245] Train net output #136: loss3/loss12 = 0.382769 (* 0.0909091 = 0.0347972 loss) | |
| I0510 16:36:32.142263 10926 solver.cpp:245] Train net output #137: loss3/loss13 = 0.107941 (* 0.0909091 = 0.00981278 loss) | |
| I0510 16:36:32.142277 10926 solver.cpp:245] Train net output #138: loss3/loss14 = 0.070714 (* 0.0909091 = 0.00642855 loss) | |
| I0510 16:36:32.142292 10926 solver.cpp:245] Train net output #139: loss3/loss15 = 0.041996 (* 0.0909091 = 0.00381782 loss) | |
| I0510 16:36:32.142305 10926 solver.cpp:245] Train net output #140: loss3/loss16 = 0.0178466 (* 0.0909091 = 0.00162242 loss) | |
| I0510 16:36:32.142319 10926 solver.cpp:245] Train net output #141: loss3/loss17 = 0.0148287 (* 0.0909091 = 0.00134806 loss) | |
| I0510 16:36:32.142333 10926 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00734161 (* 0.0909091 = 0.000667419 loss) | |
| I0510 16:36:32.142348 10926 solver.cpp:245] Train net output #143: loss3/loss19 = 0.0034245 (* 0.0909091 = 0.000311318 loss) | |
| I0510 16:36:32.142361 10926 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00170071 (* 0.0909091 = 0.00015461 loss) | |
| I0510 16:36:32.142375 10926 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00287153 (* 0.0909091 = 0.000261048 loss) | |
| I0510 16:36:32.142390 10926 solver.cpp:245] Train net output #146: loss3/loss22 = 0.0010623 (* 0.0909091 = 9.65729e-05 loss) | |
| I0510 16:36:32.142401 10926 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0510 16:36:32.142413 10926 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0510 16:36:32.142426 10926 solver.cpp:245] Train net output #149: total_confidence = 0.000108147 | |
| I0510 16:36:32.142447 10926 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.000176762 | |
| I0510 16:36:32.142462 10926 sgd_solver.cpp:106] Iteration 14000, lr = 0.001 | |
| I0510 16:37:22.998347 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 36.3513 > 30) by scale factor 0.825281 | |
| I0510 16:37:28.920202 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 30.0487 > 30) by scale factor 0.99838 | |
| I0510 16:37:30.688624 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 38.933 > 30) by scale factor 0.770555 | |
| I0510 16:38:23.226174 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 32.4358 > 30) by scale factor 0.924905 | |
| I0510 16:38:48.700212 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 33.6493 > 30) by scale factor 0.891548 | |
| I0510 16:38:59.541170 10926 solver.cpp:229] Iteration 14500, loss = 10.0832 | |
| I0510 16:38:59.541301 10926 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0363636 | |
| I0510 16:38:59.541322 10926 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.375 | |
| I0510 16:38:59.541337 10926 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.125 | |
| I0510 16:38:59.541349 10926 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0510 16:38:59.541362 10926 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.125 | |
| I0510 16:38:59.541374 10926 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.375 | |
| I0510 16:38:59.541388 10926 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.25 | |
| I0510 16:38:59.541400 10926 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.75 | |
| I0510 16:38:59.541412 10926 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.75 | |
| I0510 16:38:59.541425 10926 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.75 | |
| I0510 16:38:59.541437 10926 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.75 | |
| I0510 16:38:59.541450 10926 solver.cpp:245] Train net output #11: loss1/accuracy11 = 0.75 | |
| I0510 16:38:59.541463 10926 solver.cpp:245] Train net output #12: loss1/accuracy12 = 0.75 | |
| I0510 16:38:59.541476 10926 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0510 16:38:59.541493 10926 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0510 16:38:59.541504 10926 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0510 16:38:59.541517 10926 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0510 16:38:59.541528 10926 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0510 16:38:59.541540 10926 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0510 16:38:59.541551 10926 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0510 16:38:59.541566 10926 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0510 16:38:59.541577 10926 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0510 16:38:59.541589 10926 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0510 16:38:59.541601 10926 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.698864 | |
| I0510 16:38:59.541613 10926 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.181818 | |
| I0510 16:38:59.541630 10926 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.37459 (* 0.3 = 1.01238 loss) | |
| I0510 16:38:59.541645 10926 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.1153 (* 0.3 = 0.33459 loss) | |
| I0510 16:38:59.541659 10926 solver.cpp:245] Train net output #27: loss1/loss01 = 3.1796 (* 0.0272727 = 0.0867164 loss) | |
| I0510 16:38:59.541673 10926 solver.cpp:245] Train net output #28: loss1/loss02 = 3.39953 (* 0.0272727 = 0.0927145 loss) | |
| I0510 16:38:59.541687 10926 solver.cpp:245] Train net output #29: loss1/loss03 = 3.45746 (* 0.0272727 = 0.0942944 loss) | |
| I0510 16:38:59.541702 10926 solver.cpp:245] Train net output #30: loss1/loss04 = 2.69408 (* 0.0272727 = 0.073475 loss) | |
| I0510 16:38:59.541715 10926 solver.cpp:245] Train net output #31: loss1/loss05 = 2.46943 (* 0.0272727 = 0.067348 loss) | |
| I0510 16:38:59.541729 10926 solver.cpp:245] Train net output #32: loss1/loss06 = 3.13745 (* 0.0272727 = 0.0855668 loss) | |
| I0510 16:38:59.541744 10926 solver.cpp:245] Train net output #33: loss1/loss07 = 1.23383 (* 0.0272727 = 0.0336498 loss) | |
| I0510 16:38:59.541759 10926 solver.cpp:245] Train net output #34: loss1/loss08 = 1.06444 (* 0.0272727 = 0.0290301 loss) | |
| I0510 16:38:59.541772 10926 solver.cpp:245] Train net output #35: loss1/loss09 = 1.04682 (* 0.0272727 = 0.0285497 loss) | |
| I0510 16:38:59.541786 10926 solver.cpp:245] Train net output #36: loss1/loss10 = 0.939316 (* 0.0272727 = 0.0256177 loss) | |
| I0510 16:38:59.541800 10926 solver.cpp:245] Train net output #37: loss1/loss11 = 1.40999 (* 0.0272727 = 0.0384543 loss) | |
| I0510 16:38:59.541815 10926 solver.cpp:245] Train net output #38: loss1/loss12 = 1.13406 (* 0.0272727 = 0.030929 loss) | |
| I0510 16:38:59.541837 10926 solver.cpp:245] Train net output #39: loss1/loss13 = 0.038702 (* 0.0272727 = 0.00105551 loss) | |
| I0510 16:38:59.541870 10926 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0394677 (* 0.0272727 = 0.00107639 loss) | |
| I0510 16:38:59.541894 10926 solver.cpp:245] Train net output #41: loss1/loss15 = 0.026752 (* 0.0272727 = 0.000729599 loss) | |
| I0510 16:38:59.541909 10926 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0128343 (* 0.0272727 = 0.000350026 loss) | |
| I0510 16:38:59.541923 10926 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0145903 (* 0.0272727 = 0.000397918 loss) | |
| I0510 16:38:59.541939 10926 solver.cpp:245] Train net output #44: loss1/loss18 = 0.00456188 (* 0.0272727 = 0.000124415 loss) | |
| I0510 16:38:59.541952 10926 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0108457 (* 0.0272727 = 0.00029579 loss) | |
| I0510 16:38:59.541966 10926 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00427923 (* 0.0272727 = 0.000116706 loss) | |
| I0510 16:38:59.541980 10926 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00476789 (* 0.0272727 = 0.000130033 loss) | |
| I0510 16:38:59.541996 10926 solver.cpp:245] Train net output #48: loss1/loss22 = 0.00365446 (* 0.0272727 = 9.96671e-05 loss) | |
| I0510 16:38:59.542008 10926 solver.cpp:245] Train net output #49: loss2/accuracy = 0.127273 | |
| I0510 16:38:59.542021 10926 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.125 | |
| I0510 16:38:59.542033 10926 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0.125 | |
| I0510 16:38:59.542045 10926 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0510 16:38:59.542057 10926 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.125 | |
| I0510 16:38:59.542068 10926 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.25 | |
| I0510 16:38:59.542080 10926 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.25 | |
| I0510 16:38:59.542093 10926 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.75 | |
| I0510 16:38:59.542104 10926 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.75 | |
| I0510 16:38:59.542116 10926 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.75 | |
| I0510 16:38:59.542129 10926 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.75 | |
| I0510 16:38:59.542140 10926 solver.cpp:245] Train net output #60: loss2/accuracy11 = 0.75 | |
| I0510 16:38:59.542152 10926 solver.cpp:245] Train net output #61: loss2/accuracy12 = 0.75 | |
| I0510 16:38:59.542170 10926 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0510 16:38:59.542181 10926 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0510 16:38:59.542192 10926 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0510 16:38:59.542204 10926 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0510 16:38:59.542217 10926 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0510 16:38:59.542227 10926 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0510 16:38:59.542243 10926 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0510 16:38:59.542253 10926 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0510 16:38:59.542265 10926 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0510 16:38:59.542276 10926 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0510 16:38:59.542289 10926 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.727273 | |
| I0510 16:38:59.542300 10926 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.236364 | |
| I0510 16:38:59.542327 10926 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.51525 (* 0.3 = 1.05457 loss) | |
| I0510 16:38:59.542346 10926 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.16751 (* 0.3 = 0.350254 loss) | |
| I0510 16:38:59.542362 10926 solver.cpp:245] Train net output #76: loss2/loss01 = 3.23438 (* 0.0272727 = 0.0882103 loss) | |
| I0510 16:38:59.542385 10926 solver.cpp:245] Train net output #77: loss2/loss02 = 3.54741 (* 0.0272727 = 0.0967476 loss) | |
| I0510 16:38:59.542412 10926 solver.cpp:245] Train net output #78: loss2/loss03 = 3.29936 (* 0.0272727 = 0.0899825 loss) | |
| I0510 16:38:59.542428 10926 solver.cpp:245] Train net output #79: loss2/loss04 = 3.18119 (* 0.0272727 = 0.0867596 loss) | |
| I0510 16:38:59.542441 10926 solver.cpp:245] Train net output #80: loss2/loss05 = 2.59674 (* 0.0272727 = 0.07082 loss) | |
| I0510 16:38:59.542456 10926 solver.cpp:245] Train net output #81: loss2/loss06 = 3.11089 (* 0.0272727 = 0.0848426 loss) | |
| I0510 16:38:59.542470 10926 solver.cpp:245] Train net output #82: loss2/loss07 = 1.19675 (* 0.0272727 = 0.0326388 loss) | |
| I0510 16:38:59.542484 10926 solver.cpp:245] Train net output #83: loss2/loss08 = 1.25235 (* 0.0272727 = 0.0341551 loss) | |
| I0510 16:38:59.542505 10926 solver.cpp:245] Train net output #84: loss2/loss09 = 1.23557 (* 0.0272727 = 0.0336973 loss) | |
| I0510 16:38:59.542518 10926 solver.cpp:245] Train net output #85: loss2/loss10 = 1.04494 (* 0.0272727 = 0.0284985 loss) | |
| I0510 16:38:59.542532 10926 solver.cpp:245] Train net output #86: loss2/loss11 = 1.28786 (* 0.0272727 = 0.0351234 loss) | |
| I0510 16:38:59.542546 10926 solver.cpp:245] Train net output #87: loss2/loss12 = 1.32944 (* 0.0272727 = 0.0362574 loss) | |
| I0510 16:38:59.542559 10926 solver.cpp:245] Train net output #88: loss2/loss13 = 0.0313252 (* 0.0272727 = 0.000854323 loss) | |
| I0510 16:38:59.542573 10926 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0471177 (* 0.0272727 = 0.00128503 loss) | |
| I0510 16:38:59.542595 10926 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0427291 (* 0.0272727 = 0.00116534 loss) | |
| I0510 16:38:59.542609 10926 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0550168 (* 0.0272727 = 0.00150046 loss) | |
| I0510 16:38:59.542623 10926 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0121315 (* 0.0272727 = 0.000330858 loss) | |
| I0510 16:38:59.542639 10926 solver.cpp:245] Train net output #93: loss2/loss18 = 0.008428 (* 0.0272727 = 0.000229855 loss) | |
| I0510 16:38:59.542659 10926 solver.cpp:245] Train net output #94: loss2/loss19 = 0.0122364 (* 0.0272727 = 0.000333721 loss) | |
| I0510 16:38:59.542672 10926 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00537258 (* 0.0272727 = 0.000146525 loss) | |
| I0510 16:38:59.542686 10926 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00469705 (* 0.0272727 = 0.000128101 loss) | |
| I0510 16:38:59.542701 10926 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00350716 (* 0.0272727 = 9.56497e-05 loss) | |
| I0510 16:38:59.542713 10926 solver.cpp:245] Train net output #98: loss3/accuracy = 0.109091 | |
| I0510 16:38:59.542726 10926 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.25 | |
| I0510 16:38:59.542737 10926 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.125 | |
| I0510 16:38:59.542749 10926 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0510 16:38:59.542762 10926 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.25 | |
| I0510 16:38:59.542774 10926 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.375 | |
| I0510 16:38:59.542786 10926 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.25 | |
| I0510 16:38:59.542794 10926 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.75 | |
| I0510 16:38:59.542803 10926 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.75 | |
| I0510 16:38:59.542809 10926 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.75 | |
| I0510 16:38:59.542822 10926 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.75 | |
| I0510 16:38:59.542834 10926 solver.cpp:245] Train net output #109: loss3/accuracy11 = 0.75 | |
| I0510 16:38:59.542846 10926 solver.cpp:245] Train net output #110: loss3/accuracy12 = 0.75 | |
| I0510 16:38:59.542857 10926 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0510 16:38:59.542870 10926 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0510 16:38:59.542881 10926 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0510 16:38:59.542892 10926 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0510 16:38:59.542913 10926 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0510 16:38:59.542929 10926 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0510 16:38:59.542943 10926 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0510 16:38:59.542954 10926 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0510 16:38:59.542965 10926 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0510 16:38:59.542978 10926 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0510 16:38:59.542989 10926 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.721591 | |
| I0510 16:38:59.543000 10926 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.218182 | |
| I0510 16:38:59.543014 10926 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.38929 (* 1 = 3.38929 loss) | |
| I0510 16:38:59.543027 10926 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.09258 (* 1 = 1.09258 loss) | |
| I0510 16:38:59.543041 10926 solver.cpp:245] Train net output #125: loss3/loss01 = 3.25064 (* 0.0909091 = 0.295513 loss) | |
| I0510 16:38:59.543056 10926 solver.cpp:245] Train net output #126: loss3/loss02 = 3.57753 (* 0.0909091 = 0.32523 loss) | |
| I0510 16:38:59.543069 10926 solver.cpp:245] Train net output #127: loss3/loss03 = 3.04758 (* 0.0909091 = 0.277053 loss) | |
| I0510 16:38:59.543082 10926 solver.cpp:245] Train net output #128: loss3/loss04 = 2.6008 (* 0.0909091 = 0.236436 loss) | |
| I0510 16:38:59.543097 10926 solver.cpp:245] Train net output #129: loss3/loss05 = 2.23873 (* 0.0909091 = 0.203521 loss) | |
| I0510 16:38:59.543110 10926 solver.cpp:245] Train net output #130: loss3/loss06 = 2.88724 (* 0.0909091 = 0.262476 loss) | |
| I0510 16:38:59.543123 10926 solver.cpp:245] Train net output #131: loss3/loss07 = 1.34883 (* 0.0909091 = 0.122621 loss) | |
| I0510 16:38:59.543138 10926 solver.cpp:245] Train net output #132: loss3/loss08 = 1.43066 (* 0.0909091 = 0.13006 loss) | |
| I0510 16:38:59.543151 10926 solver.cpp:245] Train net output #133: loss3/loss09 = 0.996306 (* 0.0909091 = 0.0905733 loss) | |
| I0510 16:38:59.543165 10926 solver.cpp:245] Train net output #134: loss3/loss10 = 1.12009 (* 0.0909091 = 0.101827 loss) | |
| I0510 16:38:59.543179 10926 solver.cpp:245] Train net output #135: loss3/loss11 = 1.22301 (* 0.0909091 = 0.111183 loss) | |
| I0510 16:38:59.543193 10926 solver.cpp:245] Train net output #136: loss3/loss12 = 1.19347 (* 0.0909091 = 0.108497 loss) | |
| I0510 16:38:59.543207 10926 solver.cpp:245] Train net output #137: loss3/loss13 = 0.041963 (* 0.0909091 = 0.00381482 loss) | |
| I0510 16:38:59.543220 10926 solver.cpp:245] Train net output #138: loss3/loss14 = 0.0288085 (* 0.0909091 = 0.00261896 loss) | |
| I0510 16:38:59.543234 10926 solver.cpp:245] Train net output #139: loss3/loss15 = 0.0147387 (* 0.0909091 = 0.00133988 loss) | |
| I0510 16:38:59.543248 10926 solver.cpp:245] Train net output #140: loss3/loss16 = 0.0130395 (* 0.0909091 = 0.00118541 loss) | |
| I0510 16:38:59.543262 10926 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00559587 (* 0.0909091 = 0.000508716 loss) | |
| I0510 16:38:59.543277 10926 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00466801 (* 0.0909091 = 0.000424364 loss) | |
| I0510 16:38:59.543290 10926 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00543519 (* 0.0909091 = 0.000494108 loss) | |
| I0510 16:38:59.543304 10926 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00196241 (* 0.0909091 = 0.000178401 loss) | |
| I0510 16:38:59.543318 10926 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00164681 (* 0.0909091 = 0.00014971 loss) | |
| I0510 16:38:59.543332 10926 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00162112 (* 0.0909091 = 0.000147374 loss) | |
| I0510 16:38:59.543344 10926 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0510 16:38:59.543356 10926 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0510 16:38:59.543372 10926 solver.cpp:245] Train net output #149: total_confidence = 0.00032034 | |
| I0510 16:38:59.543395 10926 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.00152321 | |
| I0510 16:38:59.543411 10926 sgd_solver.cpp:106] Iteration 14500, lr = 0.001 | |
| I0510 16:39:41.672091 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 35.8296 > 30) by scale factor 0.837296 | |
| I0510 16:41:00.144886 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 39.6176 > 30) by scale factor 0.757239 | |
| I0510 16:41:27.279057 10926 solver.cpp:338] Iteration 15000, Testing net (#0) | |
| I0510 16:42:10.403959 10926 solver.cpp:393] Test loss: 9.37508 | |
| I0510 16:42:10.404083 10926 solver.cpp:406] Test net output #0: loss1/accuracy = 0.062556 | |
| I0510 16:42:10.404103 10926 solver.cpp:406] Test net output #1: loss1/accuracy01 = 0.12 | |
| I0510 16:42:10.404116 10926 solver.cpp:406] Test net output #2: loss1/accuracy02 = 0.089 | |
| I0510 16:42:10.404129 10926 solver.cpp:406] Test net output #3: loss1/accuracy03 = 0.065 | |
| I0510 16:42:10.404141 10926 solver.cpp:406] Test net output #4: loss1/accuracy04 = 0.163 | |
| I0510 16:42:10.404155 10926 solver.cpp:406] Test net output #5: loss1/accuracy05 = 0.302 | |
| I0510 16:42:10.404167 10926 solver.cpp:406] Test net output #6: loss1/accuracy06 = 0.455 | |
| I0510 16:42:10.404180 10926 solver.cpp:406] Test net output #7: loss1/accuracy07 = 0.732 | |
| I0510 16:42:10.404192 10926 solver.cpp:406] Test net output #8: loss1/accuracy08 = 0.913 | |
| I0510 16:42:10.404204 10926 solver.cpp:406] Test net output #9: loss1/accuracy09 = 0.989 | |
| I0510 16:42:10.404217 10926 solver.cpp:406] Test net output #10: loss1/accuracy10 = 0.999 | |
| I0510 16:42:10.404229 10926 solver.cpp:406] Test net output #11: loss1/accuracy11 = 1 | |
| I0510 16:42:10.404242 10926 solver.cpp:406] Test net output #12: loss1/accuracy12 = 1 | |
| I0510 16:42:10.404254 10926 solver.cpp:406] Test net output #13: loss1/accuracy13 = 1 | |
| I0510 16:42:10.404266 10926 solver.cpp:406] Test net output #14: loss1/accuracy14 = 1 | |
| I0510 16:42:10.404278 10926 solver.cpp:406] Test net output #15: loss1/accuracy15 = 1 | |
| I0510 16:42:10.404289 10926 solver.cpp:406] Test net output #16: loss1/accuracy16 = 1 | |
| I0510 16:42:10.404300 10926 solver.cpp:406] Test net output #17: loss1/accuracy17 = 1 | |
| I0510 16:42:10.404311 10926 solver.cpp:406] Test net output #18: loss1/accuracy18 = 1 | |
| I0510 16:42:10.404323 10926 solver.cpp:406] Test net output #19: loss1/accuracy19 = 1 | |
| I0510 16:42:10.404335 10926 solver.cpp:406] Test net output #20: loss1/accuracy20 = 1 | |
| I0510 16:42:10.404345 10926 solver.cpp:406] Test net output #21: loss1/accuracy21 = 1 | |
| I0510 16:42:10.404357 10926 solver.cpp:406] Test net output #22: loss1/accuracy22 = 1 | |
| I0510 16:42:10.404368 10926 solver.cpp:406] Test net output #23: loss1/accuracy_incl_empty = 0.763728 | |
| I0510 16:42:10.404381 10926 solver.cpp:406] Test net output #24: loss1/accuracy_top3 = 0.221731 | |
| I0510 16:42:10.404405 10926 solver.cpp:406] Test net output #25: loss1/cross_entropy_loss = 3.7472 (* 0.3 = 1.12416 loss) | |
| I0510 16:42:10.404422 10926 solver.cpp:406] Test net output #26: loss1/cross_entropy_loss_incl_empty = 0.968224 (* 0.3 = 0.290467 loss) | |
| I0510 16:42:10.404435 10926 solver.cpp:406] Test net output #27: loss1/loss01 = 3.06736 (* 0.0272727 = 0.0836553 loss) | |
| I0510 16:42:10.404449 10926 solver.cpp:406] Test net output #28: loss1/loss02 = 3.20756 (* 0.0272727 = 0.0874788 loss) | |
| I0510 16:42:10.404464 10926 solver.cpp:406] Test net output #29: loss1/loss03 = 3.30451 (* 0.0272727 = 0.090123 loss) | |
| I0510 16:42:10.404477 10926 solver.cpp:406] Test net output #30: loss1/loss04 = 3.12915 (* 0.0272727 = 0.0853406 loss) | |
| I0510 16:42:10.404490 10926 solver.cpp:406] Test net output #31: loss1/loss05 = 2.66218 (* 0.0272727 = 0.072605 loss) | |
| I0510 16:42:10.404505 10926 solver.cpp:406] Test net output #32: loss1/loss06 = 2.27617 (* 0.0272727 = 0.0620774 loss) | |
| I0510 16:42:10.404518 10926 solver.cpp:406] Test net output #33: loss1/loss07 = 1.38173 (* 0.0272727 = 0.0376837 loss) | |
| I0510 16:42:10.404532 10926 solver.cpp:406] Test net output #34: loss1/loss08 = 0.503587 (* 0.0272727 = 0.0137342 loss) | |
| I0510 16:42:10.404546 10926 solver.cpp:406] Test net output #35: loss1/loss09 = 0.0915691 (* 0.0272727 = 0.00249734 loss) | |
| I0510 16:42:10.404561 10926 solver.cpp:406] Test net output #36: loss1/loss10 = 0.0298059 (* 0.0272727 = 0.000812888 loss) | |
| I0510 16:42:10.404575 10926 solver.cpp:406] Test net output #37: loss1/loss11 = 0.0184539 (* 0.0272727 = 0.000503288 loss) | |
| I0510 16:42:10.404589 10926 solver.cpp:406] Test net output #38: loss1/loss12 = 0.012343 (* 0.0272727 = 0.000336628 loss) | |
| I0510 16:42:10.404603 10926 solver.cpp:406] Test net output #39: loss1/loss13 = 0.0105207 (* 0.0272727 = 0.000286929 loss) | |
| I0510 16:42:10.404638 10926 solver.cpp:406] Test net output #40: loss1/loss14 = 0.00785376 (* 0.0272727 = 0.000214194 loss) | |
| I0510 16:42:10.404654 10926 solver.cpp:406] Test net output #41: loss1/loss15 = 0.00571525 (* 0.0272727 = 0.00015587 loss) | |
| I0510 16:42:10.404667 10926 solver.cpp:406] Test net output #42: loss1/loss16 = 0.00443975 (* 0.0272727 = 0.000121084 loss) | |
| I0510 16:42:10.404681 10926 solver.cpp:406] Test net output #43: loss1/loss17 = 0.00298856 (* 0.0272727 = 8.15062e-05 loss) | |
| I0510 16:42:10.404695 10926 solver.cpp:406] Test net output #44: loss1/loss18 = 0.00281877 (* 0.0272727 = 7.68756e-05 loss) | |
| I0510 16:42:10.404709 10926 solver.cpp:406] Test net output #45: loss1/loss19 = 0.00242417 (* 0.0272727 = 6.61137e-05 loss) | |
| I0510 16:42:10.404724 10926 solver.cpp:406] Test net output #46: loss1/loss20 = 0.00228703 (* 0.0272727 = 6.23736e-05 loss) | |
| I0510 16:42:10.404738 10926 solver.cpp:406] Test net output #47: loss1/loss21 = 0.00207338 (* 0.0272727 = 5.65468e-05 loss) | |
| I0510 16:42:10.404752 10926 solver.cpp:406] Test net output #48: loss1/loss22 = 0.00193544 (* 0.0272727 = 5.27847e-05 loss) | |
| I0510 16:42:10.404763 10926 solver.cpp:406] Test net output #49: loss2/accuracy = 0.0700044 | |
| I0510 16:42:10.404775 10926 solver.cpp:406] Test net output #50: loss2/accuracy01 = 0.122 | |
| I0510 16:42:10.404788 10926 solver.cpp:406] Test net output #51: loss2/accuracy02 = 0.084 | |
| I0510 16:42:10.404800 10926 solver.cpp:406] Test net output #52: loss2/accuracy03 = 0.07 | |
| I0510 16:42:10.404811 10926 solver.cpp:406] Test net output #53: loss2/accuracy04 = 0.174 | |
| I0510 16:42:10.404824 10926 solver.cpp:406] Test net output #54: loss2/accuracy05 = 0.31 | |
| I0510 16:42:10.404834 10926 solver.cpp:406] Test net output #55: loss2/accuracy06 = 0.454 | |
| I0510 16:42:10.404846 10926 solver.cpp:406] Test net output #56: loss2/accuracy07 = 0.731 | |
| I0510 16:42:10.404858 10926 solver.cpp:406] Test net output #57: loss2/accuracy08 = 0.912 | |
| I0510 16:42:10.404870 10926 solver.cpp:406] Test net output #58: loss2/accuracy09 = 0.988 | |
| I0510 16:42:10.404886 10926 solver.cpp:406] Test net output #59: loss2/accuracy10 = 0.999 | |
| I0510 16:42:10.404901 10926 solver.cpp:406] Test net output #60: loss2/accuracy11 = 1 | |
| I0510 16:42:10.404908 10926 solver.cpp:406] Test net output #61: loss2/accuracy12 = 1 | |
| I0510 16:42:10.404919 10926 solver.cpp:406] Test net output #62: loss2/accuracy13 = 1 | |
| I0510 16:42:10.404932 10926 solver.cpp:406] Test net output #63: loss2/accuracy14 = 1 | |
| I0510 16:42:10.404943 10926 solver.cpp:406] Test net output #64: loss2/accuracy15 = 1 | |
| I0510 16:42:10.404954 10926 solver.cpp:406] Test net output #65: loss2/accuracy16 = 1 | |
| I0510 16:42:10.404965 10926 solver.cpp:406] Test net output #66: loss2/accuracy17 = 1 | |
| I0510 16:42:10.404983 10926 solver.cpp:406] Test net output #67: loss2/accuracy18 = 1 | |
| I0510 16:42:10.404995 10926 solver.cpp:406] Test net output #68: loss2/accuracy19 = 1 | |
| I0510 16:42:10.405006 10926 solver.cpp:406] Test net output #69: loss2/accuracy20 = 1 | |
| I0510 16:42:10.405017 10926 solver.cpp:406] Test net output #70: loss2/accuracy21 = 1 | |
| I0510 16:42:10.405030 10926 solver.cpp:406] Test net output #71: loss2/accuracy22 = 1 | |
| I0510 16:42:10.405040 10926 solver.cpp:406] Test net output #72: loss2/accuracy_incl_empty = 0.765183 | |
| I0510 16:42:10.405052 10926 solver.cpp:406] Test net output #73: loss2/accuracy_top3 = 0.230977 | |
| I0510 16:42:10.405066 10926 solver.cpp:406] Test net output #74: loss2/cross_entropy_loss = 3.59311 (* 0.3 = 1.07793 loss) | |
| I0510 16:42:10.405079 10926 solver.cpp:406] Test net output #75: loss2/cross_entropy_loss_incl_empty = 0.935389 (* 0.3 = 0.280617 loss) | |
| I0510 16:42:10.405107 10926 solver.cpp:406] Test net output #76: loss2/loss01 = 3.03319 (* 0.0272727 = 0.0827232 loss) | |
| I0510 16:42:10.405123 10926 solver.cpp:406] Test net output #77: loss2/loss02 = 3.15175 (* 0.0272727 = 0.0859569 loss) | |
| I0510 16:42:10.405138 10926 solver.cpp:406] Test net output #78: loss2/loss03 = 3.25742 (* 0.0272727 = 0.0888387 loss) | |
| I0510 16:42:10.405165 10926 solver.cpp:406] Test net output #79: loss2/loss04 = 3.08412 (* 0.0272727 = 0.0841125 loss) | |
| I0510 16:42:10.405180 10926 solver.cpp:406] Test net output #80: loss2/loss05 = 2.61143 (* 0.0272727 = 0.0712207 loss) | |
| I0510 16:42:10.405194 10926 solver.cpp:406] Test net output #81: loss2/loss06 = 2.2252 (* 0.0272727 = 0.0606872 loss) | |
| I0510 16:42:10.405208 10926 solver.cpp:406] Test net output #82: loss2/loss07 = 1.34927 (* 0.0272727 = 0.0367983 loss) | |
| I0510 16:42:10.405221 10926 solver.cpp:406] Test net output #83: loss2/loss08 = 0.503881 (* 0.0272727 = 0.0137422 loss) | |
| I0510 16:42:10.405236 10926 solver.cpp:406] Test net output #84: loss2/loss09 = 0.0886807 (* 0.0272727 = 0.00241856 loss) | |
| I0510 16:42:10.405251 10926 solver.cpp:406] Test net output #85: loss2/loss10 = 0.022737 (* 0.0272727 = 0.0006201 loss) | |
| I0510 16:42:10.405264 10926 solver.cpp:406] Test net output #86: loss2/loss11 = 0.0136235 (* 0.0272727 = 0.00037155 loss) | |
| I0510 16:42:10.405278 10926 solver.cpp:406] Test net output #87: loss2/loss12 = 0.0102671 (* 0.0272727 = 0.000280011 loss) | |
| I0510 16:42:10.405292 10926 solver.cpp:406] Test net output #88: loss2/loss13 = 0.00766583 (* 0.0272727 = 0.000209068 loss) | |
| I0510 16:42:10.405306 10926 solver.cpp:406] Test net output #89: loss2/loss14 = 0.00553581 (* 0.0272727 = 0.000150976 loss) | |
| I0510 16:42:10.405319 10926 solver.cpp:406] Test net output #90: loss2/loss15 = 0.00475144 (* 0.0272727 = 0.000129585 loss) | |
| I0510 16:42:10.405333 10926 solver.cpp:406] Test net output #91: loss2/loss16 = 0.00419866 (* 0.0272727 = 0.000114509 loss) | |
| I0510 16:42:10.405346 10926 solver.cpp:406] Test net output #92: loss2/loss17 = 0.00246574 (* 0.0272727 = 6.72475e-05 loss) | |
| I0510 16:42:10.405360 10926 solver.cpp:406] Test net output #93: loss2/loss18 = 0.00246057 (* 0.0272727 = 6.71065e-05 loss) | |
| I0510 16:42:10.405374 10926 solver.cpp:406] Test net output #94: loss2/loss19 = 0.0024441 (* 0.0272727 = 6.66574e-05 loss) | |
| I0510 16:42:10.405386 10926 solver.cpp:406] Test net output #95: loss2/loss20 = 0.00218155 (* 0.0272727 = 5.94969e-05 loss) | |
| I0510 16:42:10.405400 10926 solver.cpp:406] Test net output #96: loss2/loss21 = 0.00193337 (* 0.0272727 = 5.27283e-05 loss) | |
| I0510 16:42:10.405414 10926 solver.cpp:406] Test net output #97: loss2/loss22 = 0.0022063 (* 0.0272727 = 6.01719e-05 loss) | |
| I0510 16:42:10.405426 10926 solver.cpp:406] Test net output #98: loss3/accuracy = 0.0887496 | |
| I0510 16:42:10.405438 10926 solver.cpp:406] Test net output #99: loss3/accuracy01 = 0.125 | |
| I0510 16:42:10.405450 10926 solver.cpp:406] Test net output #100: loss3/accuracy02 = 0.101 | |
| I0510 16:42:10.405462 10926 solver.cpp:406] Test net output #101: loss3/accuracy03 = 0.076 | |
| I0510 16:42:10.405473 10926 solver.cpp:406] Test net output #102: loss3/accuracy04 = 0.175 | |
| I0510 16:42:10.405485 10926 solver.cpp:406] Test net output #103: loss3/accuracy05 = 0.318 | |
| I0510 16:42:10.405496 10926 solver.cpp:406] Test net output #104: loss3/accuracy06 = 0.448 | |
| I0510 16:42:10.405508 10926 solver.cpp:406] Test net output #105: loss3/accuracy07 = 0.729 | |
| I0510 16:42:10.405519 10926 solver.cpp:406] Test net output #106: loss3/accuracy08 = 0.911 | |
| I0510 16:42:10.405530 10926 solver.cpp:406] Test net output #107: loss3/accuracy09 = 0.988 | |
| I0510 16:42:10.405542 10926 solver.cpp:406] Test net output #108: loss3/accuracy10 = 0.999 | |
| I0510 16:42:10.405553 10926 solver.cpp:406] Test net output #109: loss3/accuracy11 = 1 | |
| I0510 16:42:10.405565 10926 solver.cpp:406] Test net output #110: loss3/accuracy12 = 1 | |
| I0510 16:42:10.405575 10926 solver.cpp:406] Test net output #111: loss3/accuracy13 = 1 | |
| I0510 16:42:10.405586 10926 solver.cpp:406] Test net output #112: loss3/accuracy14 = 1 | |
| I0510 16:42:10.405598 10926 solver.cpp:406] Test net output #113: loss3/accuracy15 = 1 | |
| I0510 16:42:10.405609 10926 solver.cpp:406] Test net output #114: loss3/accuracy16 = 1 | |
| I0510 16:42:10.405621 10926 solver.cpp:406] Test net output #115: loss3/accuracy17 = 1 | |
| I0510 16:42:10.405642 10926 solver.cpp:406] Test net output #116: loss3/accuracy18 = 1 | |
| I0510 16:42:10.405654 10926 solver.cpp:406] Test net output #117: loss3/accuracy19 = 1 | |
| I0510 16:42:10.405665 10926 solver.cpp:406] Test net output #118: loss3/accuracy20 = 1 | |
| I0510 16:42:10.405676 10926 solver.cpp:406] Test net output #119: loss3/accuracy21 = 1 | |
| I0510 16:42:10.405688 10926 solver.cpp:406] Test net output #120: loss3/accuracy22 = 1 | |
| I0510 16:42:10.405699 10926 solver.cpp:406] Test net output #121: loss3/accuracy_incl_empty = 0.764819 | |
| I0510 16:42:10.405710 10926 solver.cpp:406] Test net output #122: loss3/accuracy_top3 = 0.258436 | |
| I0510 16:42:10.405725 10926 solver.cpp:406] Test net output #123: loss3/cross_entropy_loss = 3.04755 (* 1 = 3.04755 loss) | |
| I0510 16:42:10.405737 10926 solver.cpp:406] Test net output #124: loss3/cross_entropy_loss_incl_empty = 0.837822 (* 1 = 0.837822 loss) | |
| I0510 16:42:10.405751 10926 solver.cpp:406] Test net output #125: loss3/loss01 = 2.83539 (* 0.0909091 = 0.257763 loss) | |
| I0510 16:42:10.405766 10926 solver.cpp:406] Test net output #126: loss3/loss02 = 2.98805 (* 0.0909091 = 0.271641 loss) | |
| I0510 16:42:10.405778 10926 solver.cpp:406] Test net output #127: loss3/loss03 = 3.0875 (* 0.0909091 = 0.280682 loss) | |
| I0510 16:42:10.405792 10926 solver.cpp:406] Test net output #128: loss3/loss04 = 2.92169 (* 0.0909091 = 0.265608 loss) | |
| I0510 16:42:10.405805 10926 solver.cpp:406] Test net output #129: loss3/loss05 = 2.41073 (* 0.0909091 = 0.219157 loss) | |
| I0510 16:42:10.405819 10926 solver.cpp:406] Test net output #130: loss3/loss06 = 2.05586 (* 0.0909091 = 0.186897 loss) | |
| I0510 16:42:10.405833 10926 solver.cpp:406] Test net output #131: loss3/loss07 = 1.2301 (* 0.0909091 = 0.111828 loss) | |
| I0510 16:42:10.405846 10926 solver.cpp:406] Test net output #132: loss3/loss08 = 0.461804 (* 0.0909091 = 0.0419821 loss) | |
| I0510 16:42:10.405860 10926 solver.cpp:406] Test net output #133: loss3/loss09 = 0.0892323 (* 0.0909091 = 0.00811203 loss) | |
| I0510 16:42:10.405874 10926 solver.cpp:406] Test net output #134: loss3/loss10 = 0.0226355 (* 0.0909091 = 0.00205777 loss) | |
| I0510 16:42:10.405889 10926 solver.cpp:406] Test net output #135: loss3/loss11 = 0.0142457 (* 0.0909091 = 0.00129506 loss) | |
| I0510 16:42:10.405902 10926 solver.cpp:406] Test net output #136: loss3/loss12 = 0.00923487 (* 0.0909091 = 0.000839534 loss) | |
| I0510 16:42:10.405915 10926 solver.cpp:406] Test net output #137: loss3/loss13 = 0.00618671 (* 0.0909091 = 0.000562428 loss) | |
| I0510 16:42:10.405932 10926 solver.cpp:406] Test net output #138: loss3/loss14 = 0.00367365 (* 0.0909091 = 0.000333968 loss) | |
| I0510 16:42:10.405947 10926 solver.cpp:406] Test net output #139: loss3/loss15 = 0.00267636 (* 0.0909091 = 0.000243305 loss) | |
| I0510 16:42:10.405961 10926 solver.cpp:406] Test net output #140: loss3/loss16 = 0.00195379 (* 0.0909091 = 0.000177617 loss) | |
| I0510 16:42:10.405974 10926 solver.cpp:406] Test net output #141: loss3/loss17 = 0.00136874 (* 0.0909091 = 0.000124431 loss) | |
| I0510 16:42:10.405988 10926 solver.cpp:406] Test net output #142: loss3/loss18 = 0.00127196 (* 0.0909091 = 0.000115632 loss) | |
| I0510 16:42:10.406002 10926 solver.cpp:406] Test net output #143: loss3/loss19 = 0.00111645 (* 0.0909091 = 0.000101495 loss) | |
| I0510 16:42:10.406015 10926 solver.cpp:406] Test net output #144: loss3/loss20 = 0.00100703 (* 0.0909091 = 9.15481e-05 loss) | |
| I0510 16:42:10.406029 10926 solver.cpp:406] Test net output #145: loss3/loss21 = 0.000815355 (* 0.0909091 = 7.41232e-05 loss) | |
| I0510 16:42:10.406044 10926 solver.cpp:406] Test net output #146: loss3/loss22 = 0.00075269 (* 0.0909091 = 6.84263e-05 loss) | |
| I0510 16:42:10.406055 10926 solver.cpp:406] Test net output #147: total_accuracy = 0 | |
| I0510 16:42:10.406066 10926 solver.cpp:406] Test net output #148: total_accuracy_not_rec = 0 | |
| I0510 16:42:10.406077 10926 solver.cpp:406] Test net output #149: total_confidence = 9.45527e-05 | |
| I0510 16:42:10.406090 10926 solver.cpp:406] Test net output #150: total_confidence_not_rec = 0.000257933 | |
| I0510 16:42:10.406112 10926 solver.cpp:338] Iteration 15000, Testing net (#1) | |
| I0510 16:42:53.384122 10926 solver.cpp:393] Test loss: 10.1082 | |
| I0510 16:42:53.384227 10926 solver.cpp:406] Test net output #0: loss1/accuracy = 0.0592792 | |
| I0510 16:42:53.384246 10926 solver.cpp:406] Test net output #1: loss1/accuracy01 = 0.114 | |
| I0510 16:42:53.384260 10926 solver.cpp:406] Test net output #2: loss1/accuracy02 = 0.093 | |
| I0510 16:42:53.384274 10926 solver.cpp:406] Test net output #3: loss1/accuracy03 = 0.084 | |
| I0510 16:42:53.384287 10926 solver.cpp:406] Test net output #4: loss1/accuracy04 = 0.15 | |
| I0510 16:42:53.384299 10926 solver.cpp:406] Test net output #5: loss1/accuracy05 = 0.285 | |
| I0510 16:42:53.384312 10926 solver.cpp:406] Test net output #6: loss1/accuracy06 = 0.406 | |
| I0510 16:42:53.384325 10926 solver.cpp:406] Test net output #7: loss1/accuracy07 = 0.661 | |
| I0510 16:42:53.384336 10926 solver.cpp:406] Test net output #8: loss1/accuracy08 = 0.811 | |
| I0510 16:42:53.384349 10926 solver.cpp:406] Test net output #9: loss1/accuracy09 = 0.897 | |
| I0510 16:42:53.384361 10926 solver.cpp:406] Test net output #10: loss1/accuracy10 = 0.921 | |
| I0510 16:42:53.384374 10926 solver.cpp:406] Test net output #11: loss1/accuracy11 = 0.938 | |
| I0510 16:42:53.384385 10926 solver.cpp:406] Test net output #12: loss1/accuracy12 = 0.946 | |
| I0510 16:42:53.384398 10926 solver.cpp:406] Test net output #13: loss1/accuracy13 = 0.959 | |
| I0510 16:42:53.384410 10926 solver.cpp:406] Test net output #14: loss1/accuracy14 = 0.97 | |
| I0510 16:42:53.384421 10926 solver.cpp:406] Test net output #15: loss1/accuracy15 = 0.977 | |
| I0510 16:42:53.384433 10926 solver.cpp:406] Test net output #16: loss1/accuracy16 = 0.985 | |
| I0510 16:42:53.384445 10926 solver.cpp:406] Test net output #17: loss1/accuracy17 = 0.994 | |
| I0510 16:42:53.384457 10926 solver.cpp:406] Test net output #18: loss1/accuracy18 = 0.996 | |
| I0510 16:42:53.384469 10926 solver.cpp:406] Test net output #19: loss1/accuracy19 = 0.997 | |
| I0510 16:42:53.384482 10926 solver.cpp:406] Test net output #20: loss1/accuracy20 = 0.997 | |
| I0510 16:42:53.384495 10926 solver.cpp:406] Test net output #21: loss1/accuracy21 = 0.998 | |
| I0510 16:42:53.384506 10926 solver.cpp:406] Test net output #22: loss1/accuracy22 = 0.999 | |
| I0510 16:42:53.384518 10926 solver.cpp:406] Test net output #23: loss1/accuracy_incl_empty = 0.732046 | |
| I0510 16:42:53.384531 10926 solver.cpp:406] Test net output #24: loss1/accuracy_top3 = 0.226954 | |
| I0510 16:42:53.384546 10926 solver.cpp:406] Test net output #25: loss1/cross_entropy_loss = 3.78757 (* 0.3 = 1.13627 loss) | |
| I0510 16:42:53.384562 10926 solver.cpp:406] Test net output #26: loss1/cross_entropy_loss_incl_empty = 1.11585 (* 0.3 = 0.334756 loss) | |
| I0510 16:42:53.384577 10926 solver.cpp:406] Test net output #27: loss1/loss01 = 3.08233 (* 0.0272727 = 0.0840636 loss) | |
| I0510 16:42:53.384590 10926 solver.cpp:406] Test net output #28: loss1/loss02 = 3.21483 (* 0.0272727 = 0.0876773 loss) | |
| I0510 16:42:53.384604 10926 solver.cpp:406] Test net output #29: loss1/loss03 = 3.35024 (* 0.0272727 = 0.0913702 loss) | |
| I0510 16:42:53.384618 10926 solver.cpp:406] Test net output #30: loss1/loss04 = 3.24695 (* 0.0272727 = 0.0885533 loss) | |
| I0510 16:42:53.384631 10926 solver.cpp:406] Test net output #31: loss1/loss05 = 2.81953 (* 0.0272727 = 0.0768963 loss) | |
| I0510 16:42:53.384645 10926 solver.cpp:406] Test net output #32: loss1/loss06 = 2.43736 (* 0.0272727 = 0.0664735 loss) | |
| I0510 16:42:53.384660 10926 solver.cpp:406] Test net output #33: loss1/loss07 = 1.5931 (* 0.0272727 = 0.0434482 loss) | |
| I0510 16:42:53.384672 10926 solver.cpp:406] Test net output #34: loss1/loss08 = 0.896122 (* 0.0272727 = 0.0244397 loss) | |
| I0510 16:42:53.384686 10926 solver.cpp:406] Test net output #35: loss1/loss09 = 0.455373 (* 0.0272727 = 0.0124193 loss) | |
| I0510 16:42:53.384701 10926 solver.cpp:406] Test net output #36: loss1/loss10 = 0.379327 (* 0.0272727 = 0.0103453 loss) | |
| I0510 16:42:53.384714 10926 solver.cpp:406] Test net output #37: loss1/loss11 = 0.309498 (* 0.0272727 = 0.00844085 loss) | |
| I0510 16:42:53.384728 10926 solver.cpp:406] Test net output #38: loss1/loss12 = 0.284969 (* 0.0272727 = 0.00777189 loss) | |
| I0510 16:42:53.384762 10926 solver.cpp:406] Test net output #39: loss1/loss13 = 0.224769 (* 0.0272727 = 0.00613006 loss) | |
| I0510 16:42:53.384776 10926 solver.cpp:406] Test net output #40: loss1/loss14 = 0.180043 (* 0.0272727 = 0.00491025 loss) | |
| I0510 16:42:53.384790 10926 solver.cpp:406] Test net output #41: loss1/loss15 = 0.142551 (* 0.0272727 = 0.00388774 loss) | |
| I0510 16:42:53.384804 10926 solver.cpp:406] Test net output #42: loss1/loss16 = 0.10557 (* 0.0272727 = 0.00287917 loss) | |
| I0510 16:42:53.384819 10926 solver.cpp:406] Test net output #43: loss1/loss17 = 0.047405 (* 0.0272727 = 0.00129286 loss) | |
| I0510 16:42:53.384832 10926 solver.cpp:406] Test net output #44: loss1/loss18 = 0.0357378 (* 0.0272727 = 0.000974667 loss) | |
| I0510 16:42:53.384846 10926 solver.cpp:406] Test net output #45: loss1/loss19 = 0.0276214 (* 0.0272727 = 0.000753312 loss) | |
| I0510 16:42:53.384860 10926 solver.cpp:406] Test net output #46: loss1/loss20 = 0.027903 (* 0.0272727 = 0.000760991 loss) | |
| I0510 16:42:53.384878 10926 solver.cpp:406] Test net output #47: loss1/loss21 = 0.0190969 (* 0.0272727 = 0.000520824 loss) | |
| I0510 16:42:53.384892 10926 solver.cpp:406] Test net output #48: loss1/loss22 = 0.0118652 (* 0.0272727 = 0.000323596 loss) | |
| I0510 16:42:53.384904 10926 solver.cpp:406] Test net output #49: loss2/accuracy = 0.0648979 | |
| I0510 16:42:53.384917 10926 solver.cpp:406] Test net output #50: loss2/accuracy01 = 0.117 | |
| I0510 16:42:53.384928 10926 solver.cpp:406] Test net output #51: loss2/accuracy02 = 0.086 | |
| I0510 16:42:53.384940 10926 solver.cpp:406] Test net output #52: loss2/accuracy03 = 0.074 | |
| I0510 16:42:53.384953 10926 solver.cpp:406] Test net output #53: loss2/accuracy04 = 0.149 | |
| I0510 16:42:53.384964 10926 solver.cpp:406] Test net output #54: loss2/accuracy05 = 0.29 | |
| I0510 16:42:53.384975 10926 solver.cpp:406] Test net output #55: loss2/accuracy06 = 0.407 | |
| I0510 16:42:53.384994 10926 solver.cpp:406] Test net output #56: loss2/accuracy07 = 0.657 | |
| I0510 16:42:53.385015 10926 solver.cpp:406] Test net output #57: loss2/accuracy08 = 0.812 | |
| I0510 16:42:53.385032 10926 solver.cpp:406] Test net output #58: loss2/accuracy09 = 0.896 | |
| I0510 16:42:53.385045 10926 solver.cpp:406] Test net output #59: loss2/accuracy10 = 0.921 | |
| I0510 16:42:53.385056 10926 solver.cpp:406] Test net output #60: loss2/accuracy11 = 0.938 | |
| I0510 16:42:53.385067 10926 solver.cpp:406] Test net output #61: loss2/accuracy12 = 0.946 | |
| I0510 16:42:53.385079 10926 solver.cpp:406] Test net output #62: loss2/accuracy13 = 0.959 | |
| I0510 16:42:53.385090 10926 solver.cpp:406] Test net output #63: loss2/accuracy14 = 0.97 | |
| I0510 16:42:53.385102 10926 solver.cpp:406] Test net output #64: loss2/accuracy15 = 0.977 | |
| I0510 16:42:53.385113 10926 solver.cpp:406] Test net output #65: loss2/accuracy16 = 0.985 | |
| I0510 16:42:53.385140 10926 solver.cpp:406] Test net output #66: loss2/accuracy17 = 0.994 | |
| I0510 16:42:53.385154 10926 solver.cpp:406] Test net output #67: loss2/accuracy18 = 0.996 | |
| I0510 16:42:53.385165 10926 solver.cpp:406] Test net output #68: loss2/accuracy19 = 0.997 | |
| I0510 16:42:53.385177 10926 solver.cpp:406] Test net output #69: loss2/accuracy20 = 0.997 | |
| I0510 16:42:53.385190 10926 solver.cpp:406] Test net output #70: loss2/accuracy21 = 0.998 | |
| I0510 16:42:53.385200 10926 solver.cpp:406] Test net output #71: loss2/accuracy22 = 0.999 | |
| I0510 16:42:53.385212 10926 solver.cpp:406] Test net output #72: loss2/accuracy_incl_empty = 0.732728 | |
| I0510 16:42:53.385226 10926 solver.cpp:406] Test net output #73: loss2/accuracy_top3 = 0.23319 | |
| I0510 16:42:53.385251 10926 solver.cpp:406] Test net output #74: loss2/cross_entropy_loss = 3.66995 (* 0.3 = 1.10098 loss) | |
| I0510 16:42:53.385275 10926 solver.cpp:406] Test net output #75: loss2/cross_entropy_loss_incl_empty = 1.09214 (* 0.3 = 0.327641 loss) | |
| I0510 16:42:53.385290 10926 solver.cpp:406] Test net output #76: loss2/loss01 = 3.02626 (* 0.0272727 = 0.0825345 loss) | |
| I0510 16:42:53.385305 10926 solver.cpp:406] Test net output #77: loss2/loss02 = 3.17082 (* 0.0272727 = 0.086477 loss) | |
| I0510 16:42:53.385334 10926 solver.cpp:406] Test net output #78: loss2/loss03 = 3.31392 (* 0.0272727 = 0.0903796 loss) | |
| I0510 16:42:53.385349 10926 solver.cpp:406] Test net output #79: loss2/loss04 = 3.19539 (* 0.0272727 = 0.0871471 loss) | |
| I0510 16:42:53.385362 10926 solver.cpp:406] Test net output #80: loss2/loss05 = 2.77739 (* 0.0272727 = 0.0757469 loss) | |
| I0510 16:42:53.385376 10926 solver.cpp:406] Test net output #81: loss2/loss06 = 2.39143 (* 0.0272727 = 0.0652208 loss) | |
| I0510 16:42:53.385390 10926 solver.cpp:406] Test net output #82: loss2/loss07 = 1.58046 (* 0.0272727 = 0.0431033 loss) | |
| I0510 16:42:53.385403 10926 solver.cpp:406] Test net output #83: loss2/loss08 = 0.897189 (* 0.0272727 = 0.0244688 loss) | |
| I0510 16:42:53.385417 10926 solver.cpp:406] Test net output #84: loss2/loss09 = 0.459492 (* 0.0272727 = 0.0125316 loss) | |
| I0510 16:42:53.385431 10926 solver.cpp:406] Test net output #85: loss2/loss10 = 0.369058 (* 0.0272727 = 0.0100652 loss) | |
| I0510 16:42:53.385444 10926 solver.cpp:406] Test net output #86: loss2/loss11 = 0.301369 (* 0.0272727 = 0.00821917 loss) | |
| I0510 16:42:53.385458 10926 solver.cpp:406] Test net output #87: loss2/loss12 = 0.281276 (* 0.0272727 = 0.00767117 loss) | |
| I0510 16:42:53.385473 10926 solver.cpp:406] Test net output #88: loss2/loss13 = 0.223143 (* 0.0272727 = 0.00608573 loss) | |
| I0510 16:42:53.385486 10926 solver.cpp:406] Test net output #89: loss2/loss14 = 0.176082 (* 0.0272727 = 0.00480223 loss) | |
| I0510 16:42:53.385500 10926 solver.cpp:406] Test net output #90: loss2/loss15 = 0.144825 (* 0.0272727 = 0.00394978 loss) | |
| I0510 16:42:53.385514 10926 solver.cpp:406] Test net output #91: loss2/loss16 = 0.107503 (* 0.0272727 = 0.00293189 loss) | |
| I0510 16:42:53.385529 10926 solver.cpp:406] Test net output #92: loss2/loss17 = 0.0483462 (* 0.0272727 = 0.00131853 loss) | |
| I0510 16:42:53.385541 10926 solver.cpp:406] Test net output #93: loss2/loss18 = 0.0360653 (* 0.0272727 = 0.000983599 loss) | |
| I0510 16:42:53.385555 10926 solver.cpp:406] Test net output #94: loss2/loss19 = 0.0293989 (* 0.0272727 = 0.000801789 loss) | |
| I0510 16:42:53.385570 10926 solver.cpp:406] Test net output #95: loss2/loss20 = 0.02903 (* 0.0272727 = 0.000791727 loss) | |
| I0510 16:42:53.385582 10926 solver.cpp:406] Test net output #96: loss2/loss21 = 0.0208369 (* 0.0272727 = 0.000568278 loss) | |
| I0510 16:42:53.385596 10926 solver.cpp:406] Test net output #97: loss2/loss22 = 0.0121575 (* 0.0272727 = 0.000331567 loss) | |
| I0510 16:42:53.385608 10926 solver.cpp:406] Test net output #98: loss3/accuracy = 0.0891946 | |
| I0510 16:42:53.385620 10926 solver.cpp:406] Test net output #99: loss3/accuracy01 = 0.129 | |
| I0510 16:42:53.385632 10926 solver.cpp:406] Test net output #100: loss3/accuracy02 = 0.088 | |
| I0510 16:42:53.385643 10926 solver.cpp:406] Test net output #101: loss3/accuracy03 = 0.072 | |
| I0510 16:42:53.385654 10926 solver.cpp:406] Test net output #102: loss3/accuracy04 = 0.144 | |
| I0510 16:42:53.385666 10926 solver.cpp:406] Test net output #103: loss3/accuracy05 = 0.296 | |
| I0510 16:42:53.385678 10926 solver.cpp:406] Test net output #104: loss3/accuracy06 = 0.411 | |
| I0510 16:42:53.385689 10926 solver.cpp:406] Test net output #105: loss3/accuracy07 = 0.663 | |
| I0510 16:42:53.385700 10926 solver.cpp:406] Test net output #106: loss3/accuracy08 = 0.813 | |
| I0510 16:42:53.385712 10926 solver.cpp:406] Test net output #107: loss3/accuracy09 = 0.903 | |
| I0510 16:42:53.385723 10926 solver.cpp:406] Test net output #108: loss3/accuracy10 = 0.92 | |
| I0510 16:42:53.385735 10926 solver.cpp:406] Test net output #109: loss3/accuracy11 = 0.938 | |
| I0510 16:42:53.385746 10926 solver.cpp:406] Test net output #110: loss3/accuracy12 = 0.946 | |
| I0510 16:42:53.385757 10926 solver.cpp:406] Test net output #111: loss3/accuracy13 = 0.959 | |
| I0510 16:42:53.385769 10926 solver.cpp:406] Test net output #112: loss3/accuracy14 = 0.97 | |
| I0510 16:42:53.385782 10926 solver.cpp:406] Test net output #113: loss3/accuracy15 = 0.977 | |
| I0510 16:42:53.385792 10926 solver.cpp:406] Test net output #114: loss3/accuracy16 = 0.985 | |
| I0510 16:42:53.385814 10926 solver.cpp:406] Test net output #115: loss3/accuracy17 = 0.994 | |
| I0510 16:42:53.385828 10926 solver.cpp:406] Test net output #116: loss3/accuracy18 = 0.996 | |
| I0510 16:42:53.385839 10926 solver.cpp:406] Test net output #117: loss3/accuracy19 = 0.997 | |
| I0510 16:42:53.385851 10926 solver.cpp:406] Test net output #118: loss3/accuracy20 = 0.997 | |
| I0510 16:42:53.385862 10926 solver.cpp:406] Test net output #119: loss3/accuracy21 = 0.998 | |
| I0510 16:42:53.385874 10926 solver.cpp:406] Test net output #120: loss3/accuracy22 = 0.999 | |
| I0510 16:42:53.385885 10926 solver.cpp:406] Test net output #121: loss3/accuracy_incl_empty = 0.734411 | |
| I0510 16:42:53.385897 10926 solver.cpp:406] Test net output #122: loss3/accuracy_top3 = 0.257493 | |
| I0510 16:42:53.385911 10926 solver.cpp:406] Test net output #123: loss3/cross_entropy_loss = 3.08608 (* 1 = 3.08608 loss) | |
| I0510 16:42:53.385927 10926 solver.cpp:406] Test net output #124: loss3/cross_entropy_loss_incl_empty = 0.965787 (* 1 = 0.965787 loss) | |
| I0510 16:42:53.385942 10926 solver.cpp:406] Test net output #125: loss3/loss01 = 2.83547 (* 0.0909091 = 0.25777 loss) | |
| I0510 16:42:53.385957 10926 solver.cpp:406] Test net output #126: loss3/loss02 = 3.01519 (* 0.0909091 = 0.274109 loss) | |
| I0510 16:42:53.385969 10926 solver.cpp:406] Test net output #127: loss3/loss03 = 3.10939 (* 0.0909091 = 0.282672 loss) | |
| I0510 16:42:53.385983 10926 solver.cpp:406] Test net output #128: loss3/loss04 = 3.01364 (* 0.0909091 = 0.273968 loss) | |
| I0510 16:42:53.385996 10926 solver.cpp:406] Test net output #129: loss3/loss05 = 2.573 (* 0.0909091 = 0.233909 loss) | |
| I0510 16:42:53.386010 10926 solver.cpp:406] Test net output #130: loss3/loss06 = 2.23987 (* 0.0909091 = 0.203624 loss) | |
| I0510 16:42:53.386023 10926 solver.cpp:406] Test net output #131: loss3/loss07 = 1.44584 (* 0.0909091 = 0.13144 loss) | |
| I0510 16:42:53.386037 10926 solver.cpp:406] Test net output #132: loss3/loss08 = 0.832906 (* 0.0909091 = 0.0757188 loss) | |
| I0510 16:42:53.386050 10926 solver.cpp:406] Test net output #133: loss3/loss09 = 0.423598 (* 0.0909091 = 0.0385089 loss) | |
| I0510 16:42:53.386065 10926 solver.cpp:406] Test net output #134: loss3/loss10 = 0.342057 (* 0.0909091 = 0.0310961 loss) | |
| I0510 16:42:53.386081 10926 solver.cpp:406] Test net output #135: loss3/loss11 = 0.273364 (* 0.0909091 = 0.0248513 loss) | |
| I0510 16:42:53.386096 10926 solver.cpp:406] Test net output #136: loss3/loss12 = 0.252389 (* 0.0909091 = 0.0229445 loss) | |
| I0510 16:42:53.386108 10926 solver.cpp:406] Test net output #137: loss3/loss13 = 0.197892 (* 0.0909091 = 0.0179902 loss) | |
| I0510 16:42:53.386122 10926 solver.cpp:406] Test net output #138: loss3/loss14 = 0.149824 (* 0.0909091 = 0.0136204 loss) | |
| I0510 16:42:53.386135 10926 solver.cpp:406] Test net output #139: loss3/loss15 = 0.12123 (* 0.0909091 = 0.011021 loss) | |
| I0510 16:42:53.386149 10926 solver.cpp:406] Test net output #140: loss3/loss16 = 0.0912763 (* 0.0909091 = 0.00829784 loss) | |
| I0510 16:42:53.386162 10926 solver.cpp:406] Test net output #141: loss3/loss17 = 0.0428221 (* 0.0909091 = 0.00389291 loss) | |
| I0510 16:42:53.386176 10926 solver.cpp:406] Test net output #142: loss3/loss18 = 0.0333679 (* 0.0909091 = 0.00303344 loss) | |
| I0510 16:42:53.386190 10926 solver.cpp:406] Test net output #143: loss3/loss19 = 0.0261453 (* 0.0909091 = 0.00237685 loss) | |
| I0510 16:42:53.386204 10926 solver.cpp:406] Test net output #144: loss3/loss20 = 0.0287573 (* 0.0909091 = 0.0026143 loss) | |
| I0510 16:42:53.386217 10926 solver.cpp:406] Test net output #145: loss3/loss21 = 0.0188439 (* 0.0909091 = 0.00171308 loss) | |
| I0510 16:42:53.386231 10926 solver.cpp:406] Test net output #146: loss3/loss22 = 0.0115564 (* 0.0909091 = 0.00105059 loss) | |
| I0510 16:42:53.386243 10926 solver.cpp:406] Test net output #147: total_accuracy = 0 | |
| I0510 16:42:53.386255 10926 solver.cpp:406] Test net output #148: total_accuracy_not_rec = 0 | |
| I0510 16:42:53.386265 10926 solver.cpp:406] Test net output #149: total_confidence = 7.72697e-05 | |
| I0510 16:42:53.386287 10926 solver.cpp:406] Test net output #150: total_confidence_not_rec = 0.000223368 | |
| I0510 16:42:53.531936 10926 solver.cpp:229] Iteration 15000, loss = 9.95834 | |
| I0510 16:42:53.532012 10926 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0576923 | |
| I0510 16:42:53.532032 10926 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.125 | |
| I0510 16:42:53.532047 10926 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.125 | |
| I0510 16:42:53.532059 10926 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.125 | |
| I0510 16:42:53.532073 10926 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.25 | |
| I0510 16:42:53.532084 10926 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.375 | |
| I0510 16:42:53.532097 10926 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.375 | |
| I0510 16:42:53.532110 10926 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.5 | |
| I0510 16:42:53.532124 10926 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.75 | |
| I0510 16:42:53.532135 10926 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.875 | |
| I0510 16:42:53.532148 10926 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.875 | |
| I0510 16:42:53.532162 10926 solver.cpp:245] Train net output #11: loss1/accuracy11 = 0.875 | |
| I0510 16:42:53.532174 10926 solver.cpp:245] Train net output #12: loss1/accuracy12 = 0.875 | |
| I0510 16:42:53.532187 10926 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0510 16:42:53.532199 10926 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0510 16:42:53.532212 10926 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0510 16:42:53.532224 10926 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0510 16:42:53.532237 10926 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0510 16:42:53.532248 10926 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0510 16:42:53.532260 10926 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0510 16:42:53.532272 10926 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0510 16:42:53.532285 10926 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0510 16:42:53.532297 10926 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0510 16:42:53.532310 10926 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.721591 | |
| I0510 16:42:53.532322 10926 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.269231 | |
| I0510 16:42:53.532340 10926 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.23857 (* 0.3 = 0.971571 loss) | |
| I0510 16:42:53.532354 10926 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.0133 (* 0.3 = 0.30399 loss) | |
| I0510 16:42:53.532368 10926 solver.cpp:245] Train net output #27: loss1/loss01 = 3.82311 (* 0.0272727 = 0.104267 loss) | |
| I0510 16:42:53.532383 10926 solver.cpp:245] Train net output #28: loss1/loss02 = 3.04526 (* 0.0272727 = 0.0830525 loss) | |
| I0510 16:42:53.532397 10926 solver.cpp:245] Train net output #29: loss1/loss03 = 3.50573 (* 0.0272727 = 0.0956109 loss) | |
| I0510 16:42:53.532413 10926 solver.cpp:245] Train net output #30: loss1/loss04 = 3.24387 (* 0.0272727 = 0.0884692 loss) | |
| I0510 16:42:53.532426 10926 solver.cpp:245] Train net output #31: loss1/loss05 = 2.43677 (* 0.0272727 = 0.0664574 loss) | |
| I0510 16:42:53.532441 10926 solver.cpp:245] Train net output #32: loss1/loss06 = 2.39109 (* 0.0272727 = 0.0652116 loss) | |
| I0510 16:42:53.532455 10926 solver.cpp:245] Train net output #33: loss1/loss07 = 1.91624 (* 0.0272727 = 0.0522612 loss) | |
| I0510 16:42:53.532469 10926 solver.cpp:245] Train net output #34: loss1/loss08 = 1.40775 (* 0.0272727 = 0.0383933 loss) | |
| I0510 16:42:53.532483 10926 solver.cpp:245] Train net output #35: loss1/loss09 = 0.592638 (* 0.0272727 = 0.0161629 loss) | |
| I0510 16:42:53.532498 10926 solver.cpp:245] Train net output #36: loss1/loss10 = 0.591918 (* 0.0272727 = 0.0161432 loss) | |
| I0510 16:42:53.532512 10926 solver.cpp:245] Train net output #37: loss1/loss11 = 0.374856 (* 0.0272727 = 0.0102233 loss) | |
| I0510 16:42:53.532562 10926 solver.cpp:245] Train net output #38: loss1/loss12 = 0.751305 (* 0.0272727 = 0.0204901 loss) | |
| I0510 16:42:53.532578 10926 solver.cpp:245] Train net output #39: loss1/loss13 = 0.0376875 (* 0.0272727 = 0.00102784 loss) | |
| I0510 16:42:53.532593 10926 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0295097 (* 0.0272727 = 0.000804811 loss) | |
| I0510 16:42:53.532608 10926 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0201575 (* 0.0272727 = 0.000549751 loss) | |
| I0510 16:42:53.532621 10926 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0157596 (* 0.0272727 = 0.000429806 loss) | |
| I0510 16:42:53.532635 10926 solver.cpp:245] Train net output #43: loss1/loss17 = 0.00731489 (* 0.0272727 = 0.000199497 loss) | |
| I0510 16:42:53.532654 10926 solver.cpp:245] Train net output #44: loss1/loss18 = 0.00762809 (* 0.0272727 = 0.000208039 loss) | |
| I0510 16:42:53.532668 10926 solver.cpp:245] Train net output #45: loss1/loss19 = 0.00662081 (* 0.0272727 = 0.000180568 loss) | |
| I0510 16:42:53.532683 10926 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00444477 (* 0.0272727 = 0.000121221 loss) | |
| I0510 16:42:53.532697 10926 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00494882 (* 0.0272727 = 0.000134968 loss) | |
| I0510 16:42:53.532712 10926 solver.cpp:245] Train net output #48: loss1/loss22 = 0.003784 (* 0.0272727 = 0.0001032 loss) | |
| I0510 16:42:53.532726 10926 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0576923 | |
| I0510 16:42:53.532737 10926 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0510 16:42:53.532749 10926 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0510 16:42:53.532763 10926 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0510 16:42:53.532773 10926 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.125 | |
| I0510 16:42:53.532785 10926 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.375 | |
| I0510 16:42:53.532798 10926 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.375 | |
| I0510 16:42:53.532814 10926 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.5 | |
| I0510 16:42:53.532827 10926 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.75 | |
| I0510 16:42:53.532840 10926 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.875 | |
| I0510 16:42:53.532851 10926 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.875 | |
| I0510 16:42:53.532863 10926 solver.cpp:245] Train net output #60: loss2/accuracy11 = 0.875 | |
| I0510 16:42:53.532876 10926 solver.cpp:245] Train net output #61: loss2/accuracy12 = 0.875 | |
| I0510 16:42:53.532888 10926 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0510 16:42:53.532901 10926 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0510 16:42:53.532912 10926 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0510 16:42:53.532924 10926 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0510 16:42:53.532937 10926 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0510 16:42:53.532948 10926 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0510 16:42:53.532959 10926 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0510 16:42:53.532971 10926 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0510 16:42:53.532984 10926 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0510 16:42:53.532995 10926 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0510 16:42:53.533007 10926 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.721591 | |
| I0510 16:42:53.533020 10926 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.25 | |
| I0510 16:42:53.533033 10926 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.2722 (* 0.3 = 0.981661 loss) | |
| I0510 16:42:53.533047 10926 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.02302 (* 0.3 = 0.306907 loss) | |
| I0510 16:42:53.533071 10926 solver.cpp:245] Train net output #76: loss2/loss01 = 3.86477 (* 0.0272727 = 0.105403 loss) | |
| I0510 16:42:53.533082 10926 solver.cpp:245] Train net output #77: loss2/loss02 = 3.44124 (* 0.0272727 = 0.0938519 loss) | |
| I0510 16:42:53.533092 10926 solver.cpp:245] Train net output #78: loss2/loss03 = 3.41487 (* 0.0272727 = 0.0931327 loss) | |
| I0510 16:42:53.533107 10926 solver.cpp:245] Train net output #79: loss2/loss04 = 3.68787 (* 0.0272727 = 0.100578 loss) | |
| I0510 16:42:53.533135 10926 solver.cpp:245] Train net output #80: loss2/loss05 = 2.37237 (* 0.0272727 = 0.0647011 loss) | |
| I0510 16:42:53.533151 10926 solver.cpp:245] Train net output #81: loss2/loss06 = 2.32723 (* 0.0272727 = 0.0634698 loss) | |
| I0510 16:42:53.533166 10926 solver.cpp:245] Train net output #82: loss2/loss07 = 2.0142 (* 0.0272727 = 0.0549328 loss) | |
| I0510 16:42:53.533180 10926 solver.cpp:245] Train net output #83: loss2/loss08 = 1.31729 (* 0.0272727 = 0.035926 loss) | |
| I0510 16:42:53.533195 10926 solver.cpp:245] Train net output #84: loss2/loss09 = 0.617416 (* 0.0272727 = 0.0168386 loss) | |
| I0510 16:42:53.533210 10926 solver.cpp:245] Train net output #85: loss2/loss10 = 0.604797 (* 0.0272727 = 0.0164945 loss) | |
| I0510 16:42:53.533223 10926 solver.cpp:245] Train net output #86: loss2/loss11 = 0.6238 (* 0.0272727 = 0.0170127 loss) | |
| I0510 16:42:53.533237 10926 solver.cpp:245] Train net output #87: loss2/loss12 = 0.702283 (* 0.0272727 = 0.0191532 loss) | |
| I0510 16:42:53.533252 10926 solver.cpp:245] Train net output #88: loss2/loss13 = 0.0450606 (* 0.0272727 = 0.00122893 loss) | |
| I0510 16:42:53.533267 10926 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0465539 (* 0.0272727 = 0.00126965 loss) | |
| I0510 16:42:53.533282 10926 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0195843 (* 0.0272727 = 0.000534116 loss) | |
| I0510 16:42:53.533295 10926 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0333366 (* 0.0272727 = 0.000909179 loss) | |
| I0510 16:42:53.533309 10926 solver.cpp:245] Train net output #92: loss2/loss17 = 0.00931449 (* 0.0272727 = 0.000254032 loss) | |
| I0510 16:42:53.533324 10926 solver.cpp:245] Train net output #93: loss2/loss18 = 0.00795802 (* 0.0272727 = 0.000217037 loss) | |
| I0510 16:42:53.533337 10926 solver.cpp:245] Train net output #94: loss2/loss19 = 0.00390675 (* 0.0272727 = 0.000106548 loss) | |
| I0510 16:42:53.533351 10926 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00680415 (* 0.0272727 = 0.000185568 loss) | |
| I0510 16:42:53.533366 10926 solver.cpp:245] Train net output #96: loss2/loss21 = 0.0112893 (* 0.0272727 = 0.00030789 loss) | |
| I0510 16:42:53.533380 10926 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00306081 (* 0.0272727 = 8.34766e-05 loss) | |
| I0510 16:42:53.533393 10926 solver.cpp:245] Train net output #98: loss3/accuracy = 0.134615 | |
| I0510 16:42:53.533406 10926 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0510 16:42:53.533418 10926 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0510 16:42:53.533429 10926 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0.125 | |
| I0510 16:42:53.533442 10926 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0510 16:42:53.533453 10926 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.375 | |
| I0510 16:42:53.533466 10926 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.5 | |
| I0510 16:42:53.533478 10926 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.5 | |
| I0510 16:42:53.533490 10926 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.75 | |
| I0510 16:42:53.533502 10926 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.875 | |
| I0510 16:42:53.533514 10926 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.875 | |
| I0510 16:42:53.533526 10926 solver.cpp:245] Train net output #109: loss3/accuracy11 = 0.875 | |
| I0510 16:42:53.533538 10926 solver.cpp:245] Train net output #110: loss3/accuracy12 = 0.875 | |
| I0510 16:42:53.533550 10926 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0510 16:42:53.533574 10926 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0510 16:42:53.533587 10926 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0510 16:42:53.533601 10926 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0510 16:42:53.533612 10926 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0510 16:42:53.533624 10926 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0510 16:42:53.533635 10926 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0510 16:42:53.533648 10926 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0510 16:42:53.533659 10926 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0510 16:42:53.533671 10926 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0510 16:42:53.533684 10926 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.732955 | |
| I0510 16:42:53.533695 10926 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.288462 | |
| I0510 16:42:53.533712 10926 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.04448 (* 1 = 3.04448 loss) | |
| I0510 16:42:53.533727 10926 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 0.995719 (* 1 = 0.995719 loss) | |
| I0510 16:42:53.533742 10926 solver.cpp:245] Train net output #125: loss3/loss01 = 3.68471 (* 0.0909091 = 0.334974 loss) | |
| I0510 16:42:53.533756 10926 solver.cpp:245] Train net output #126: loss3/loss02 = 2.77171 (* 0.0909091 = 0.251973 loss) | |
| I0510 16:42:53.533771 10926 solver.cpp:245] Train net output #127: loss3/loss03 = 3.06519 (* 0.0909091 = 0.278654 loss) | |
| I0510 16:42:53.533785 10926 solver.cpp:245] Train net output #128: loss3/loss04 = 3.04822 (* 0.0909091 = 0.277111 loss) | |
| I0510 16:42:53.533799 10926 solver.cpp:245] Train net output #129: loss3/loss05 = 2.16174 (* 0.0909091 = 0.196522 loss) | |
| I0510 16:42:53.533814 10926 solver.cpp:245] Train net output #130: loss3/loss06 = 1.81208 (* 0.0909091 = 0.164734 loss) | |
| I0510 16:42:53.533828 10926 solver.cpp:245] Train net output #131: loss3/loss07 = 1.59679 (* 0.0909091 = 0.145163 loss) | |
| I0510 16:42:53.533841 10926 solver.cpp:245] Train net output #132: loss3/loss08 = 0.995562 (* 0.0909091 = 0.0905056 loss) | |
| I0510 16:42:53.533859 10926 solver.cpp:245] Train net output #133: loss3/loss09 = 0.448409 (* 0.0909091 = 0.0407644 loss) | |
| I0510 16:42:53.533874 10926 solver.cpp:245] Train net output #134: loss3/loss10 = 0.464137 (* 0.0909091 = 0.0421943 loss) | |
| I0510 16:42:53.533888 10926 solver.cpp:245] Train net output #135: loss3/loss11 = 0.401905 (* 0.0909091 = 0.0365368 loss) | |
| I0510 16:42:53.533903 10926 solver.cpp:245] Train net output #136: loss3/loss12 = 0.666341 (* 0.0909091 = 0.0605764 loss) | |
| I0510 16:42:53.533917 10926 solver.cpp:245] Train net output #137: loss3/loss13 = 0.0208566 (* 0.0909091 = 0.00189606 loss) | |
| I0510 16:42:53.533932 10926 solver.cpp:245] Train net output #138: loss3/loss14 = 0.0123364 (* 0.0909091 = 0.00112149 loss) | |
| I0510 16:42:53.533947 10926 solver.cpp:245] Train net output #139: loss3/loss15 = 0.00905564 (* 0.0909091 = 0.00082324 loss) | |
| I0510 16:42:53.533962 10926 solver.cpp:245] Train net output #140: loss3/loss16 = 0.00420058 (* 0.0909091 = 0.000381871 loss) | |
| I0510 16:42:53.533975 10926 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00514203 (* 0.0909091 = 0.000467458 loss) | |
| I0510 16:42:53.533990 10926 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00442323 (* 0.0909091 = 0.000402112 loss) | |
| I0510 16:42:53.534004 10926 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00264544 (* 0.0909091 = 0.000240494 loss) | |
| I0510 16:42:53.534018 10926 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00179822 (* 0.0909091 = 0.000163474 loss) | |
| I0510 16:42:53.534034 10926 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00164519 (* 0.0909091 = 0.000149563 loss) | |
| I0510 16:42:53.534049 10926 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00125076 (* 0.0909091 = 0.000113705 loss) | |
| I0510 16:42:53.534070 10926 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0510 16:42:53.534085 10926 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0510 16:42:53.534096 10926 solver.cpp:245] Train net output #149: total_confidence = 2.57877e-05 | |
| I0510 16:42:53.534108 10926 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.000306176 | |
| I0510 16:42:53.534121 10926 sgd_solver.cpp:106] Iteration 15000, lr = 0.001 | |
| I0510 16:43:15.981914 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 32.455 > 30) by scale factor 0.924358 | |
| I0510 16:44:19.342597 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 30.6121 > 30) by scale factor 0.980005 | |
| I0510 16:44:49.706835 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 38.4614 > 30) by scale factor 0.780004 | |
| I0510 16:45:20.802850 10926 solver.cpp:229] Iteration 15500, loss = 10.0331 | |
| I0510 16:45:20.802968 10926 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0819672 | |
| I0510 16:45:20.802989 10926 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0510 16:45:20.803004 10926 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.125 | |
| I0510 16:45:20.803017 10926 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0510 16:45:20.803030 10926 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0 | |
| I0510 16:45:20.803041 10926 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.125 | |
| I0510 16:45:20.803055 10926 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.125 | |
| I0510 16:45:20.803068 10926 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.25 | |
| I0510 16:45:20.803081 10926 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.375 | |
| I0510 16:45:20.803093 10926 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.75 | |
| I0510 16:45:20.803105 10926 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.75 | |
| I0510 16:45:20.803118 10926 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0510 16:45:20.803131 10926 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0510 16:45:20.803143 10926 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0510 16:45:20.803155 10926 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0510 16:45:20.803167 10926 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0510 16:45:20.803179 10926 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0510 16:45:20.803191 10926 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0510 16:45:20.803203 10926 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0510 16:45:20.803215 10926 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0510 16:45:20.803227 10926 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0510 16:45:20.803239 10926 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0510 16:45:20.803251 10926 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0510 16:45:20.803263 10926 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.659091 | |
| I0510 16:45:20.803275 10926 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.180328 | |
| I0510 16:45:20.803292 10926 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.30057 (* 0.3 = 0.99017 loss) | |
| I0510 16:45:20.803308 10926 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.35911 (* 0.3 = 0.407732 loss) | |
| I0510 16:45:20.803323 10926 solver.cpp:245] Train net output #27: loss1/loss01 = 3.41768 (* 0.0272727 = 0.0932093 loss) | |
| I0510 16:45:20.803336 10926 solver.cpp:245] Train net output #28: loss1/loss02 = 3.95569 (* 0.0272727 = 0.107882 loss) | |
| I0510 16:45:20.803352 10926 solver.cpp:245] Train net output #29: loss1/loss03 = 3.41978 (* 0.0272727 = 0.0932667 loss) | |
| I0510 16:45:20.803366 10926 solver.cpp:245] Train net output #30: loss1/loss04 = 3.5841 (* 0.0272727 = 0.0977481 loss) | |
| I0510 16:45:20.803380 10926 solver.cpp:245] Train net output #31: loss1/loss05 = 3.2735 (* 0.0272727 = 0.0892772 loss) | |
| I0510 16:45:20.803395 10926 solver.cpp:245] Train net output #32: loss1/loss06 = 3.38934 (* 0.0272727 = 0.0924365 loss) | |
| I0510 16:45:20.803408 10926 solver.cpp:245] Train net output #33: loss1/loss07 = 2.94217 (* 0.0272727 = 0.080241 loss) | |
| I0510 16:45:20.803422 10926 solver.cpp:245] Train net output #34: loss1/loss08 = 3.02079 (* 0.0272727 = 0.0823853 loss) | |
| I0510 16:45:20.803437 10926 solver.cpp:245] Train net output #35: loss1/loss09 = 1.16501 (* 0.0272727 = 0.0317729 loss) | |
| I0510 16:45:20.803452 10926 solver.cpp:245] Train net output #36: loss1/loss10 = 1.60194 (* 0.0272727 = 0.0436892 loss) | |
| I0510 16:45:20.803465 10926 solver.cpp:245] Train net output #37: loss1/loss11 = 0.0898226 (* 0.0272727 = 0.00244971 loss) | |
| I0510 16:45:20.803480 10926 solver.cpp:245] Train net output #38: loss1/loss12 = 0.0660174 (* 0.0272727 = 0.00180047 loss) | |
| I0510 16:45:20.803494 10926 solver.cpp:245] Train net output #39: loss1/loss13 = 0.0586454 (* 0.0272727 = 0.00159942 loss) | |
| I0510 16:45:20.803527 10926 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0366258 (* 0.0272727 = 0.000998885 loss) | |
| I0510 16:45:20.803544 10926 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0253071 (* 0.0272727 = 0.000690194 loss) | |
| I0510 16:45:20.803558 10926 solver.cpp:245] Train net output #42: loss1/loss16 = 0.020145 (* 0.0272727 = 0.00054941 loss) | |
| I0510 16:45:20.803572 10926 solver.cpp:245] Train net output #43: loss1/loss17 = 0.00701617 (* 0.0272727 = 0.00019135 loss) | |
| I0510 16:45:20.803587 10926 solver.cpp:245] Train net output #44: loss1/loss18 = 0.00681937 (* 0.0272727 = 0.000185983 loss) | |
| I0510 16:45:20.803601 10926 solver.cpp:245] Train net output #45: loss1/loss19 = 0.00604608 (* 0.0272727 = 0.000164893 loss) | |
| I0510 16:45:20.803616 10926 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00473544 (* 0.0272727 = 0.000129148 loss) | |
| I0510 16:45:20.803629 10926 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00375183 (* 0.0272727 = 0.000102323 loss) | |
| I0510 16:45:20.803643 10926 solver.cpp:245] Train net output #48: loss1/loss22 = 0.00256045 (* 0.0272727 = 6.98303e-05 loss) | |
| I0510 16:45:20.803656 10926 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0655738 | |
| I0510 16:45:20.803668 10926 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.125 | |
| I0510 16:45:20.803680 10926 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0510 16:45:20.803692 10926 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0510 16:45:20.803704 10926 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0 | |
| I0510 16:45:20.803715 10926 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.125 | |
| I0510 16:45:20.803727 10926 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.125 | |
| I0510 16:45:20.803740 10926 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.25 | |
| I0510 16:45:20.803751 10926 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.375 | |
| I0510 16:45:20.803763 10926 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.75 | |
| I0510 16:45:20.803776 10926 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.75 | |
| I0510 16:45:20.803788 10926 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0510 16:45:20.803799 10926 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0510 16:45:20.803812 10926 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0510 16:45:20.803823 10926 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0510 16:45:20.803834 10926 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0510 16:45:20.803848 10926 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0510 16:45:20.803859 10926 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0510 16:45:20.803870 10926 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0510 16:45:20.803885 10926 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0510 16:45:20.803897 10926 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0510 16:45:20.803910 10926 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0510 16:45:20.803921 10926 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0510 16:45:20.803933 10926 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.670455 | |
| I0510 16:45:20.803946 10926 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.245902 | |
| I0510 16:45:20.803959 10926 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.2644 (* 0.3 = 0.979319 loss) | |
| I0510 16:45:20.803974 10926 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.2118 (* 0.3 = 0.36354 loss) | |
| I0510 16:45:20.803992 10926 solver.cpp:245] Train net output #76: loss2/loss01 = 3.49495 (* 0.0272727 = 0.0953168 loss) | |
| I0510 16:45:20.804006 10926 solver.cpp:245] Train net output #77: loss2/loss02 = 3.6822 (* 0.0272727 = 0.100424 loss) | |
| I0510 16:45:20.804031 10926 solver.cpp:245] Train net output #78: loss2/loss03 = 3.40322 (* 0.0272727 = 0.092815 loss) | |
| I0510 16:45:20.804047 10926 solver.cpp:245] Train net output #79: loss2/loss04 = 3.73994 (* 0.0272727 = 0.101998 loss) | |
| I0510 16:45:20.804061 10926 solver.cpp:245] Train net output #80: loss2/loss05 = 2.99277 (* 0.0272727 = 0.081621 loss) | |
| I0510 16:45:20.804075 10926 solver.cpp:245] Train net output #81: loss2/loss06 = 3.39034 (* 0.0272727 = 0.0924639 loss) | |
| I0510 16:45:20.804090 10926 solver.cpp:245] Train net output #82: loss2/loss07 = 2.9758 (* 0.0272727 = 0.0811581 loss) | |
| I0510 16:45:20.804103 10926 solver.cpp:245] Train net output #83: loss2/loss08 = 2.65421 (* 0.0272727 = 0.0723877 loss) | |
| I0510 16:45:20.804117 10926 solver.cpp:245] Train net output #84: loss2/loss09 = 1.20636 (* 0.0272727 = 0.0329006 loss) | |
| I0510 16:45:20.804131 10926 solver.cpp:245] Train net output #85: loss2/loss10 = 1.65093 (* 0.0272727 = 0.0450253 loss) | |
| I0510 16:45:20.804146 10926 solver.cpp:245] Train net output #86: loss2/loss11 = 0.130306 (* 0.0272727 = 0.00355379 loss) | |
| I0510 16:45:20.804159 10926 solver.cpp:245] Train net output #87: loss2/loss12 = 0.104362 (* 0.0272727 = 0.00284624 loss) | |
| I0510 16:45:20.804170 10926 solver.cpp:245] Train net output #88: loss2/loss13 = 0.0850298 (* 0.0272727 = 0.002319 loss) | |
| I0510 16:45:20.804180 10926 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0751451 (* 0.0272727 = 0.00204941 loss) | |
| I0510 16:45:20.804195 10926 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0588222 (* 0.0272727 = 0.00160424 loss) | |
| I0510 16:45:20.804209 10926 solver.cpp:245] Train net output #91: loss2/loss16 = 0.037944 (* 0.0272727 = 0.00103484 loss) | |
| I0510 16:45:20.804224 10926 solver.cpp:245] Train net output #92: loss2/loss17 = 0.026821 (* 0.0272727 = 0.000731483 loss) | |
| I0510 16:45:20.804239 10926 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0129907 (* 0.0272727 = 0.000354291 loss) | |
| I0510 16:45:20.804252 10926 solver.cpp:245] Train net output #94: loss2/loss19 = 0.00942565 (* 0.0272727 = 0.000257063 loss) | |
| I0510 16:45:20.804266 10926 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00694693 (* 0.0272727 = 0.000189462 loss) | |
| I0510 16:45:20.804280 10926 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00761406 (* 0.0272727 = 0.000207656 loss) | |
| I0510 16:45:20.804293 10926 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00380415 (* 0.0272727 = 0.00010375 loss) | |
| I0510 16:45:20.804306 10926 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0327869 | |
| I0510 16:45:20.804318 10926 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0510 16:45:20.804332 10926 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0510 16:45:20.804342 10926 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0.125 | |
| I0510 16:45:20.804354 10926 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0510 16:45:20.804365 10926 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.125 | |
| I0510 16:45:20.804378 10926 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.125 | |
| I0510 16:45:20.804389 10926 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.25 | |
| I0510 16:45:20.804400 10926 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.375 | |
| I0510 16:45:20.804411 10926 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.75 | |
| I0510 16:45:20.804424 10926 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.75 | |
| I0510 16:45:20.804435 10926 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0510 16:45:20.804446 10926 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0510 16:45:20.804457 10926 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0510 16:45:20.804469 10926 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0510 16:45:20.804481 10926 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0510 16:45:20.804491 10926 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0510 16:45:20.804512 10926 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0510 16:45:20.804527 10926 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0510 16:45:20.804538 10926 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0510 16:45:20.804549 10926 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0510 16:45:20.804561 10926 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0510 16:45:20.804572 10926 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0510 16:45:20.804584 10926 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.659091 | |
| I0510 16:45:20.804596 10926 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.229508 | |
| I0510 16:45:20.804610 10926 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.30842 (* 1 = 3.30842 loss) | |
| I0510 16:45:20.804623 10926 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.22054 (* 1 = 1.22054 loss) | |
| I0510 16:45:20.804637 10926 solver.cpp:245] Train net output #125: loss3/loss01 = 2.79341 (* 0.0909091 = 0.253946 loss) | |
| I0510 16:45:20.804651 10926 solver.cpp:245] Train net output #126: loss3/loss02 = 3.72245 (* 0.0909091 = 0.338405 loss) | |
| I0510 16:45:20.804666 10926 solver.cpp:245] Train net output #127: loss3/loss03 = 3.23949 (* 0.0909091 = 0.294499 loss) | |
| I0510 16:45:20.804679 10926 solver.cpp:245] Train net output #128: loss3/loss04 = 3.50786 (* 0.0909091 = 0.318896 loss) | |
| I0510 16:45:20.804693 10926 solver.cpp:245] Train net output #129: loss3/loss05 = 2.94094 (* 0.0909091 = 0.267359 loss) | |
| I0510 16:45:20.804708 10926 solver.cpp:245] Train net output #130: loss3/loss06 = 3.21425 (* 0.0909091 = 0.292205 loss) | |
| I0510 16:45:20.804721 10926 solver.cpp:245] Train net output #131: loss3/loss07 = 2.77514 (* 0.0909091 = 0.252286 loss) | |
| I0510 16:45:20.804735 10926 solver.cpp:245] Train net output #132: loss3/loss08 = 2.72629 (* 0.0909091 = 0.247845 loss) | |
| I0510 16:45:20.804749 10926 solver.cpp:245] Train net output #133: loss3/loss09 = 1.30656 (* 0.0909091 = 0.118778 loss) | |
| I0510 16:45:20.804764 10926 solver.cpp:245] Train net output #134: loss3/loss10 = 1.62908 (* 0.0909091 = 0.148098 loss) | |
| I0510 16:45:20.804777 10926 solver.cpp:245] Train net output #135: loss3/loss11 = 0.0808714 (* 0.0909091 = 0.00735195 loss) | |
| I0510 16:45:20.804791 10926 solver.cpp:245] Train net output #136: loss3/loss12 = 0.0564388 (* 0.0909091 = 0.0051308 loss) | |
| I0510 16:45:20.804806 10926 solver.cpp:245] Train net output #137: loss3/loss13 = 0.0322123 (* 0.0909091 = 0.00292839 loss) | |
| I0510 16:45:20.804821 10926 solver.cpp:245] Train net output #138: loss3/loss14 = 0.0268909 (* 0.0909091 = 0.00244463 loss) | |
| I0510 16:45:20.804834 10926 solver.cpp:245] Train net output #139: loss3/loss15 = 0.0163129 (* 0.0909091 = 0.00148299 loss) | |
| I0510 16:45:20.804847 10926 solver.cpp:245] Train net output #140: loss3/loss16 = 0.0170164 (* 0.0909091 = 0.00154695 loss) | |
| I0510 16:45:20.804862 10926 solver.cpp:245] Train net output #141: loss3/loss17 = 0.0127502 (* 0.0909091 = 0.00115911 loss) | |
| I0510 16:45:20.804875 10926 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00805804 (* 0.0909091 = 0.000732549 loss) | |
| I0510 16:45:20.804889 10926 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00634005 (* 0.0909091 = 0.000576368 loss) | |
| I0510 16:45:20.804903 10926 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00479623 (* 0.0909091 = 0.000436021 loss) | |
| I0510 16:45:20.804918 10926 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00289768 (* 0.0909091 = 0.000263425 loss) | |
| I0510 16:45:20.804934 10926 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00151736 (* 0.0909091 = 0.000137941 loss) | |
| I0510 16:45:20.804947 10926 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0510 16:45:20.804960 10926 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0510 16:45:20.804971 10926 solver.cpp:245] Train net output #149: total_confidence = 2.93085e-07 | |
| I0510 16:45:20.804993 10926 solver.cpp:245] Train net output #150: total_confidence_not_rec = 1.9509e-06 | |
| I0510 16:45:20.805008 10926 sgd_solver.cpp:106] Iteration 15500, lr = 0.001 | |
| I0510 16:46:20.403260 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 39.3161 > 30) by scale factor 0.763046 | |
| I0510 16:46:39.814960 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 75.5931 > 30) by scale factor 0.396862 | |
| I0510 16:47:48.213774 10926 solver.cpp:229] Iteration 16000, loss = 9.91933 | |
| I0510 16:47:48.213989 10926 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0212766 | |
| I0510 16:47:48.214013 10926 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.125 | |
| I0510 16:47:48.214028 10926 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.125 | |
| I0510 16:47:48.214042 10926 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0510 16:47:48.214056 10926 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.25 | |
| I0510 16:47:48.214068 10926 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.375 | |
| I0510 16:47:48.214082 10926 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.375 | |
| I0510 16:47:48.214094 10926 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.625 | |
| I0510 16:47:48.214107 10926 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.875 | |
| I0510 16:47:48.214120 10926 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.875 | |
| I0510 16:47:48.214133 10926 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.875 | |
| I0510 16:47:48.214146 10926 solver.cpp:245] Train net output #11: loss1/accuracy11 = 0.875 | |
| I0510 16:47:48.214159 10926 solver.cpp:245] Train net output #12: loss1/accuracy12 = 0.875 | |
| I0510 16:47:48.214171 10926 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0510 16:47:48.214184 10926 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0510 16:47:48.214195 10926 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0510 16:47:48.214207 10926 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0510 16:47:48.214220 10926 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0510 16:47:48.214231 10926 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0510 16:47:48.214243 10926 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0510 16:47:48.214256 10926 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0510 16:47:48.214267 10926 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0510 16:47:48.214279 10926 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0510 16:47:48.214292 10926 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.732955 | |
| I0510 16:47:48.214303 10926 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.148936 | |
| I0510 16:47:48.214320 10926 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.24032 (* 0.3 = 0.972097 loss) | |
| I0510 16:47:48.214335 10926 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.01192 (* 0.3 = 0.303575 loss) | |
| I0510 16:47:48.214350 10926 solver.cpp:245] Train net output #27: loss1/loss01 = 3.30265 (* 0.0272727 = 0.0900723 loss) | |
| I0510 16:47:48.214365 10926 solver.cpp:245] Train net output #28: loss1/loss02 = 3.16389 (* 0.0272727 = 0.086288 loss) | |
| I0510 16:47:48.214380 10926 solver.cpp:245] Train net output #29: loss1/loss03 = 4.04773 (* 0.0272727 = 0.110393 loss) | |
| I0510 16:47:48.214395 10926 solver.cpp:245] Train net output #30: loss1/loss04 = 2.7959 (* 0.0272727 = 0.0762518 loss) | |
| I0510 16:47:48.214408 10926 solver.cpp:245] Train net output #31: loss1/loss05 = 2.25804 (* 0.0272727 = 0.0615828 loss) | |
| I0510 16:47:48.214423 10926 solver.cpp:245] Train net output #32: loss1/loss06 = 2.73713 (* 0.0272727 = 0.0746491 loss) | |
| I0510 16:47:48.214437 10926 solver.cpp:245] Train net output #33: loss1/loss07 = 1.44576 (* 0.0272727 = 0.0394297 loss) | |
| I0510 16:47:48.214452 10926 solver.cpp:245] Train net output #34: loss1/loss08 = 0.493813 (* 0.0272727 = 0.0134676 loss) | |
| I0510 16:47:48.214467 10926 solver.cpp:245] Train net output #35: loss1/loss09 = 0.671762 (* 0.0272727 = 0.0183208 loss) | |
| I0510 16:47:48.214481 10926 solver.cpp:245] Train net output #36: loss1/loss10 = 0.672783 (* 0.0272727 = 0.0183486 loss) | |
| I0510 16:47:48.214495 10926 solver.cpp:245] Train net output #37: loss1/loss11 = 0.585596 (* 0.0272727 = 0.0159708 loss) | |
| I0510 16:47:48.214510 10926 solver.cpp:245] Train net output #38: loss1/loss12 = 0.612156 (* 0.0272727 = 0.0166952 loss) | |
| I0510 16:47:48.214541 10926 solver.cpp:245] Train net output #39: loss1/loss13 = 0.116498 (* 0.0272727 = 0.00317721 loss) | |
| I0510 16:47:48.214557 10926 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0864407 (* 0.0272727 = 0.00235747 loss) | |
| I0510 16:47:48.214571 10926 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0627644 (* 0.0272727 = 0.00171176 loss) | |
| I0510 16:47:48.214586 10926 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0613735 (* 0.0272727 = 0.00167382 loss) | |
| I0510 16:47:48.214601 10926 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0216623 (* 0.0272727 = 0.00059079 loss) | |
| I0510 16:47:48.214615 10926 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0377472 (* 0.0272727 = 0.00102947 loss) | |
| I0510 16:47:48.214630 10926 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0291495 (* 0.0272727 = 0.000794987 loss) | |
| I0510 16:47:48.214645 10926 solver.cpp:245] Train net output #46: loss1/loss20 = 0.052042 (* 0.0272727 = 0.00141933 loss) | |
| I0510 16:47:48.214659 10926 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0234188 (* 0.0272727 = 0.000638695 loss) | |
| I0510 16:47:48.214673 10926 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0201338 (* 0.0272727 = 0.000549103 loss) | |
| I0510 16:47:48.214687 10926 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0212766 | |
| I0510 16:47:48.214699 10926 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.375 | |
| I0510 16:47:48.214711 10926 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0.125 | |
| I0510 16:47:48.214723 10926 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0510 16:47:48.214735 10926 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.375 | |
| I0510 16:47:48.214747 10926 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.375 | |
| I0510 16:47:48.214759 10926 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.5 | |
| I0510 16:47:48.214772 10926 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.625 | |
| I0510 16:47:48.214784 10926 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.875 | |
| I0510 16:47:48.214797 10926 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.875 | |
| I0510 16:47:48.214808 10926 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.875 | |
| I0510 16:47:48.214821 10926 solver.cpp:245] Train net output #60: loss2/accuracy11 = 0.875 | |
| I0510 16:47:48.214833 10926 solver.cpp:245] Train net output #61: loss2/accuracy12 = 0.875 | |
| I0510 16:47:48.214845 10926 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0510 16:47:48.214857 10926 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0510 16:47:48.214869 10926 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0510 16:47:48.214884 10926 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0510 16:47:48.214895 10926 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0510 16:47:48.214907 10926 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0510 16:47:48.214920 10926 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0510 16:47:48.214931 10926 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0510 16:47:48.214943 10926 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0510 16:47:48.214954 10926 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0510 16:47:48.214967 10926 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.715909 | |
| I0510 16:47:48.214977 10926 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.170213 | |
| I0510 16:47:48.214994 10926 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.26137 (* 0.3 = 0.978412 loss) | |
| I0510 16:47:48.215010 10926 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.12917 (* 0.3 = 0.33875 loss) | |
| I0510 16:47:48.215025 10926 solver.cpp:245] Train net output #76: loss2/loss01 = 3.23031 (* 0.0272727 = 0.0880993 loss) | |
| I0510 16:47:48.215039 10926 solver.cpp:245] Train net output #77: loss2/loss02 = 3.19243 (* 0.0272727 = 0.0870663 loss) | |
| I0510 16:47:48.215065 10926 solver.cpp:245] Train net output #78: loss2/loss03 = 3.58183 (* 0.0272727 = 0.0976863 loss) | |
| I0510 16:47:48.215081 10926 solver.cpp:245] Train net output #79: loss2/loss04 = 2.89867 (* 0.0272727 = 0.0790547 loss) | |
| I0510 16:47:48.215095 10926 solver.cpp:245] Train net output #80: loss2/loss05 = 2.61171 (* 0.0272727 = 0.0712285 loss) | |
| I0510 16:47:48.215109 10926 solver.cpp:245] Train net output #81: loss2/loss06 = 2.38504 (* 0.0272727 = 0.0650465 loss) | |
| I0510 16:47:48.215124 10926 solver.cpp:245] Train net output #82: loss2/loss07 = 1.60743 (* 0.0272727 = 0.0438389 loss) | |
| I0510 16:47:48.215138 10926 solver.cpp:245] Train net output #83: loss2/loss08 = 0.615459 (* 0.0272727 = 0.0167852 loss) | |
| I0510 16:47:48.215152 10926 solver.cpp:245] Train net output #84: loss2/loss09 = 0.898042 (* 0.0272727 = 0.024492 loss) | |
| I0510 16:47:48.215167 10926 solver.cpp:245] Train net output #85: loss2/loss10 = 0.847287 (* 0.0272727 = 0.0231078 loss) | |
| I0510 16:47:48.215181 10926 solver.cpp:245] Train net output #86: loss2/loss11 = 0.629053 (* 0.0272727 = 0.017156 loss) | |
| I0510 16:47:48.215195 10926 solver.cpp:245] Train net output #87: loss2/loss12 = 0.55455 (* 0.0272727 = 0.0151241 loss) | |
| I0510 16:47:48.215209 10926 solver.cpp:245] Train net output #88: loss2/loss13 = 0.121252 (* 0.0272727 = 0.00330688 loss) | |
| I0510 16:47:48.215224 10926 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0858302 (* 0.0272727 = 0.00234082 loss) | |
| I0510 16:47:48.215240 10926 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0415103 (* 0.0272727 = 0.0011321 loss) | |
| I0510 16:47:48.215255 10926 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0410892 (* 0.0272727 = 0.00112061 loss) | |
| I0510 16:47:48.215268 10926 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0163202 (* 0.0272727 = 0.000445095 loss) | |
| I0510 16:47:48.215282 10926 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0277469 (* 0.0272727 = 0.000756733 loss) | |
| I0510 16:47:48.215297 10926 solver.cpp:245] Train net output #94: loss2/loss19 = 0.023101 (* 0.0272727 = 0.000630026 loss) | |
| I0510 16:47:48.215312 10926 solver.cpp:245] Train net output #95: loss2/loss20 = 0.0283335 (* 0.0272727 = 0.000772732 loss) | |
| I0510 16:47:48.215327 10926 solver.cpp:245] Train net output #96: loss2/loss21 = 0.025153 (* 0.0272727 = 0.00068599 loss) | |
| I0510 16:47:48.215340 10926 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0185418 (* 0.0272727 = 0.000505687 loss) | |
| I0510 16:47:48.215354 10926 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0638298 | |
| I0510 16:47:48.215363 10926 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.125 | |
| I0510 16:47:48.215371 10926 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0510 16:47:48.215384 10926 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0510 16:47:48.215396 10926 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.25 | |
| I0510 16:47:48.215409 10926 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.5 | |
| I0510 16:47:48.215421 10926 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.5 | |
| I0510 16:47:48.215433 10926 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.75 | |
| I0510 16:47:48.215445 10926 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.875 | |
| I0510 16:47:48.215456 10926 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.875 | |
| I0510 16:47:48.215468 10926 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.875 | |
| I0510 16:47:48.215481 10926 solver.cpp:245] Train net output #109: loss3/accuracy11 = 0.875 | |
| I0510 16:47:48.215492 10926 solver.cpp:245] Train net output #110: loss3/accuracy12 = 0.875 | |
| I0510 16:47:48.215503 10926 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0510 16:47:48.215515 10926 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0510 16:47:48.215526 10926 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0510 16:47:48.215538 10926 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0510 16:47:48.215574 10926 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0510 16:47:48.215589 10926 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0510 16:47:48.215600 10926 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0510 16:47:48.215612 10926 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0510 16:47:48.215625 10926 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0510 16:47:48.215636 10926 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0510 16:47:48.215648 10926 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.732955 | |
| I0510 16:47:48.215661 10926 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.212766 | |
| I0510 16:47:48.215675 10926 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.05279 (* 1 = 3.05279 loss) | |
| I0510 16:47:48.215689 10926 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.00626 (* 1 = 1.00626 loss) | |
| I0510 16:47:48.215703 10926 solver.cpp:245] Train net output #125: loss3/loss01 = 2.84232 (* 0.0909091 = 0.258393 loss) | |
| I0510 16:47:48.215718 10926 solver.cpp:245] Train net output #126: loss3/loss02 = 2.97507 (* 0.0909091 = 0.270461 loss) | |
| I0510 16:47:48.215731 10926 solver.cpp:245] Train net output #127: loss3/loss03 = 3.46204 (* 0.0909091 = 0.314731 loss) | |
| I0510 16:47:48.215745 10926 solver.cpp:245] Train net output #128: loss3/loss04 = 2.59832 (* 0.0909091 = 0.236211 loss) | |
| I0510 16:47:48.215759 10926 solver.cpp:245] Train net output #129: loss3/loss05 = 1.98118 (* 0.0909091 = 0.180108 loss) | |
| I0510 16:47:48.215773 10926 solver.cpp:245] Train net output #130: loss3/loss06 = 1.98893 (* 0.0909091 = 0.180812 loss) | |
| I0510 16:47:48.215787 10926 solver.cpp:245] Train net output #131: loss3/loss07 = 0.985175 (* 0.0909091 = 0.0895614 loss) | |
| I0510 16:47:48.215801 10926 solver.cpp:245] Train net output #132: loss3/loss08 = 0.575056 (* 0.0909091 = 0.0522778 loss) | |
| I0510 16:47:48.215816 10926 solver.cpp:245] Train net output #133: loss3/loss09 = 0.765748 (* 0.0909091 = 0.0696135 loss) | |
| I0510 16:47:48.215831 10926 solver.cpp:245] Train net output #134: loss3/loss10 = 0.675217 (* 0.0909091 = 0.0613834 loss) | |
| I0510 16:47:48.215844 10926 solver.cpp:245] Train net output #135: loss3/loss11 = 0.37827 (* 0.0909091 = 0.0343882 loss) | |
| I0510 16:47:48.215859 10926 solver.cpp:245] Train net output #136: loss3/loss12 = 0.430607 (* 0.0909091 = 0.0391461 loss) | |
| I0510 16:47:48.215873 10926 solver.cpp:245] Train net output #137: loss3/loss13 = 0.123096 (* 0.0909091 = 0.0111906 loss) | |
| I0510 16:47:48.215888 10926 solver.cpp:245] Train net output #138: loss3/loss14 = 0.0576429 (* 0.0909091 = 0.00524026 loss) | |
| I0510 16:47:48.215903 10926 solver.cpp:245] Train net output #139: loss3/loss15 = 0.0539957 (* 0.0909091 = 0.0049087 loss) | |
| I0510 16:47:48.215917 10926 solver.cpp:245] Train net output #140: loss3/loss16 = 0.014677 (* 0.0909091 = 0.00133428 loss) | |
| I0510 16:47:48.215935 10926 solver.cpp:245] Train net output #141: loss3/loss17 = 0.0114367 (* 0.0909091 = 0.0010397 loss) | |
| I0510 16:47:48.215951 10926 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00783496 (* 0.0909091 = 0.000712269 loss) | |
| I0510 16:47:48.215965 10926 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00504566 (* 0.0909091 = 0.000458697 loss) | |
| I0510 16:47:48.215981 10926 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00257579 (* 0.0909091 = 0.000234163 loss) | |
| I0510 16:47:48.215994 10926 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00233912 (* 0.0909091 = 0.000212647 loss) | |
| I0510 16:47:48.216009 10926 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00129666 (* 0.0909091 = 0.000117878 loss) | |
| I0510 16:47:48.216022 10926 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0510 16:47:48.216033 10926 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0510 16:47:48.216048 10926 solver.cpp:245] Train net output #149: total_confidence = 9.7542e-05 | |
| I0510 16:47:48.216070 10926 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.000703303 | |
| I0510 16:47:48.216086 10926 sgd_solver.cpp:106] Iteration 16000, lr = 0.001 | |
| I0510 16:48:30.106066 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 55.0591 > 30) by scale factor 0.544869 | |
| I0510 16:48:37.462204 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 32.7626 > 30) by scale factor 0.915677 | |
| I0510 16:48:51.608417 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 43.582 > 30) by scale factor 0.688358 | |
| I0510 16:49:04.282274 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 30.8231 > 30) by scale factor 0.973297 | |
| I0510 16:49:07.819759 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 30.7703 > 30) by scale factor 0.974966 | |
| I0510 16:49:38.711925 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 41.551 > 30) by scale factor 0.722005 | |
| I0510 16:50:15.162662 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 112.063 > 30) by scale factor 0.267706 | |
| I0510 16:50:15.358642 10926 solver.cpp:229] Iteration 16500, loss = 9.96728 | |
| I0510 16:50:15.358716 10926 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0909091 | |
| I0510 16:50:15.358736 10926 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0510 16:50:15.358750 10926 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.25 | |
| I0510 16:50:15.358763 10926 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.125 | |
| I0510 16:50:15.358777 10926 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.125 | |
| I0510 16:50:15.358793 10926 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.375 | |
| I0510 16:50:15.358806 10926 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.375 | |
| I0510 16:50:15.358819 10926 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.625 | |
| I0510 16:50:15.358832 10926 solver.cpp:245] Train net output #8: loss1/accuracy08 = 1 | |
| I0510 16:50:15.358845 10926 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0510 16:50:15.358857 10926 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0510 16:50:15.358870 10926 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0510 16:50:15.358883 10926 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0510 16:50:15.358896 10926 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0510 16:50:15.358908 10926 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0510 16:50:15.358921 10926 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0510 16:50:15.358932 10926 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0510 16:50:15.358944 10926 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0510 16:50:15.358957 10926 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0510 16:50:15.358968 10926 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0510 16:50:15.358980 10926 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0510 16:50:15.358992 10926 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0510 16:50:15.359005 10926 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0510 16:50:15.359017 10926 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.772727 | |
| I0510 16:50:15.359030 10926 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.204545 | |
| I0510 16:50:15.359047 10926 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.29987 (* 0.3 = 0.989961 loss) | |
| I0510 16:50:15.359062 10926 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 0.940444 (* 0.3 = 0.282133 loss) | |
| I0510 16:50:15.359077 10926 solver.cpp:245] Train net output #27: loss1/loss01 = 3.05598 (* 0.0272727 = 0.0833449 loss) | |
| I0510 16:50:15.359091 10926 solver.cpp:245] Train net output #28: loss1/loss02 = 3.28196 (* 0.0272727 = 0.089508 loss) | |
| I0510 16:50:15.359105 10926 solver.cpp:245] Train net output #29: loss1/loss03 = 3.58269 (* 0.0272727 = 0.0977098 loss) | |
| I0510 16:50:15.359120 10926 solver.cpp:245] Train net output #30: loss1/loss04 = 3.62187 (* 0.0272727 = 0.0987782 loss) | |
| I0510 16:50:15.359134 10926 solver.cpp:245] Train net output #31: loss1/loss05 = 2.88253 (* 0.0272727 = 0.0786144 loss) | |
| I0510 16:50:15.359149 10926 solver.cpp:245] Train net output #32: loss1/loss06 = 2.34436 (* 0.0272727 = 0.0639371 loss) | |
| I0510 16:50:15.359163 10926 solver.cpp:245] Train net output #33: loss1/loss07 = 1.94771 (* 0.0272727 = 0.0531193 loss) | |
| I0510 16:50:15.359179 10926 solver.cpp:245] Train net output #34: loss1/loss08 = 0.105975 (* 0.0272727 = 0.00289023 loss) | |
| I0510 16:50:15.359194 10926 solver.cpp:245] Train net output #35: loss1/loss09 = 0.0219523 (* 0.0272727 = 0.000598699 loss) | |
| I0510 16:50:15.359208 10926 solver.cpp:245] Train net output #36: loss1/loss10 = 0.00803013 (* 0.0272727 = 0.000219003 loss) | |
| I0510 16:50:15.359222 10926 solver.cpp:245] Train net output #37: loss1/loss11 = 0.0123414 (* 0.0272727 = 0.000336582 loss) | |
| I0510 16:50:15.359237 10926 solver.cpp:245] Train net output #38: loss1/loss12 = 0.00557443 (* 0.0272727 = 0.00015203 loss) | |
| I0510 16:50:15.359287 10926 solver.cpp:245] Train net output #39: loss1/loss13 = 0.00565768 (* 0.0272727 = 0.0001543 loss) | |
| I0510 16:50:15.359302 10926 solver.cpp:245] Train net output #40: loss1/loss14 = 0.00527559 (* 0.0272727 = 0.00014388 loss) | |
| I0510 16:50:15.359318 10926 solver.cpp:245] Train net output #41: loss1/loss15 = 0.00165399 (* 0.0272727 = 4.51088e-05 loss) | |
| I0510 16:50:15.359333 10926 solver.cpp:245] Train net output #42: loss1/loss16 = 0.00361085 (* 0.0272727 = 9.84776e-05 loss) | |
| I0510 16:50:15.359346 10926 solver.cpp:245] Train net output #43: loss1/loss17 = 0.00172467 (* 0.0272727 = 4.70364e-05 loss) | |
| I0510 16:50:15.359360 10926 solver.cpp:245] Train net output #44: loss1/loss18 = 0.00158687 (* 0.0272727 = 4.32783e-05 loss) | |
| I0510 16:50:15.359375 10926 solver.cpp:245] Train net output #45: loss1/loss19 = 0.00512976 (* 0.0272727 = 0.000139902 loss) | |
| I0510 16:50:15.359390 10926 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00187248 (* 0.0272727 = 5.10677e-05 loss) | |
| I0510 16:50:15.359403 10926 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00205753 (* 0.0272727 = 5.61144e-05 loss) | |
| I0510 16:50:15.359418 10926 solver.cpp:245] Train net output #48: loss1/loss22 = 0.00301429 (* 0.0272727 = 8.2208e-05 loss) | |
| I0510 16:50:15.359431 10926 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0454545 | |
| I0510 16:50:15.359443 10926 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.125 | |
| I0510 16:50:15.359457 10926 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0.25 | |
| I0510 16:50:15.359468 10926 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0510 16:50:15.359482 10926 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.125 | |
| I0510 16:50:15.359493 10926 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.25 | |
| I0510 16:50:15.359505 10926 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.5 | |
| I0510 16:50:15.359519 10926 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.625 | |
| I0510 16:50:15.359530 10926 solver.cpp:245] Train net output #57: loss2/accuracy08 = 1 | |
| I0510 16:50:15.359542 10926 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0510 16:50:15.359555 10926 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0510 16:50:15.359566 10926 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0510 16:50:15.359578 10926 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0510 16:50:15.359591 10926 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0510 16:50:15.359602 10926 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0510 16:50:15.359614 10926 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0510 16:50:15.359627 10926 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0510 16:50:15.359638 10926 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0510 16:50:15.359650 10926 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0510 16:50:15.359663 10926 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0510 16:50:15.359674 10926 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0510 16:50:15.359688 10926 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0510 16:50:15.359699 10926 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0510 16:50:15.359711 10926 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.761364 | |
| I0510 16:50:15.359726 10926 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.204545 | |
| I0510 16:50:15.359743 10926 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.38533 (* 0.3 = 1.0156 loss) | |
| I0510 16:50:15.359760 10926 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 0.926885 (* 0.3 = 0.278066 loss) | |
| I0510 16:50:15.359774 10926 solver.cpp:245] Train net output #76: loss2/loss01 = 3.18602 (* 0.0272727 = 0.0868914 loss) | |
| I0510 16:50:15.359789 10926 solver.cpp:245] Train net output #77: loss2/loss02 = 3.27797 (* 0.0272727 = 0.0893993 loss) | |
| I0510 16:50:15.359815 10926 solver.cpp:245] Train net output #78: loss2/loss03 = 3.98508 (* 0.0272727 = 0.108684 loss) | |
| I0510 16:50:15.359830 10926 solver.cpp:245] Train net output #79: loss2/loss04 = 3.52829 (* 0.0272727 = 0.096226 loss) | |
| I0510 16:50:15.359849 10926 solver.cpp:245] Train net output #80: loss2/loss05 = 2.58918 (* 0.0272727 = 0.070614 loss) | |
| I0510 16:50:15.359863 10926 solver.cpp:245] Train net output #81: loss2/loss06 = 2.30192 (* 0.0272727 = 0.0627796 loss) | |
| I0510 16:50:15.359879 10926 solver.cpp:245] Train net output #82: loss2/loss07 = 2.03355 (* 0.0272727 = 0.0554605 loss) | |
| I0510 16:50:15.359892 10926 solver.cpp:245] Train net output #83: loss2/loss08 = 0.0682775 (* 0.0272727 = 0.00186211 loss) | |
| I0510 16:50:15.359907 10926 solver.cpp:245] Train net output #84: loss2/loss09 = 0.0231698 (* 0.0272727 = 0.000631903 loss) | |
| I0510 16:50:15.359922 10926 solver.cpp:245] Train net output #85: loss2/loss10 = 0.0164883 (* 0.0272727 = 0.00044968 loss) | |
| I0510 16:50:15.359936 10926 solver.cpp:245] Train net output #86: loss2/loss11 = 0.0130502 (* 0.0272727 = 0.000355914 loss) | |
| I0510 16:50:15.359951 10926 solver.cpp:245] Train net output #87: loss2/loss12 = 0.010236 (* 0.0272727 = 0.000279164 loss) | |
| I0510 16:50:15.359966 10926 solver.cpp:245] Train net output #88: loss2/loss13 = 0.0156287 (* 0.0272727 = 0.000426238 loss) | |
| I0510 16:50:15.359980 10926 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0147548 (* 0.0272727 = 0.000402405 loss) | |
| I0510 16:50:15.359994 10926 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0082974 (* 0.0272727 = 0.000226293 loss) | |
| I0510 16:50:15.360009 10926 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0136375 (* 0.0272727 = 0.000371931 loss) | |
| I0510 16:50:15.360023 10926 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0142879 (* 0.0272727 = 0.000389669 loss) | |
| I0510 16:50:15.360038 10926 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0122268 (* 0.0272727 = 0.000333459 loss) | |
| I0510 16:50:15.360052 10926 solver.cpp:245] Train net output #94: loss2/loss19 = 0.0055847 (* 0.0272727 = 0.00015231 loss) | |
| I0510 16:50:15.360066 10926 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00850951 (* 0.0272727 = 0.000232078 loss) | |
| I0510 16:50:15.360080 10926 solver.cpp:245] Train net output #96: loss2/loss21 = 0.0122326 (* 0.0272727 = 0.000333615 loss) | |
| I0510 16:50:15.360095 10926 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0124585 (* 0.0272727 = 0.000339776 loss) | |
| I0510 16:50:15.360107 10926 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0454545 | |
| I0510 16:50:15.360121 10926 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0510 16:50:15.360132 10926 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0510 16:50:15.360144 10926 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0510 16:50:15.360155 10926 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0510 16:50:15.360167 10926 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.25 | |
| I0510 16:50:15.360179 10926 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.5 | |
| I0510 16:50:15.360191 10926 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.625 | |
| I0510 16:50:15.360203 10926 solver.cpp:245] Train net output #106: loss3/accuracy08 = 1 | |
| I0510 16:50:15.360214 10926 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0510 16:50:15.360226 10926 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0510 16:50:15.360239 10926 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0510 16:50:15.360250 10926 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0510 16:50:15.360261 10926 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0510 16:50:15.360273 10926 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0510 16:50:15.360285 10926 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0510 16:50:15.360307 10926 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0510 16:50:15.360321 10926 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0510 16:50:15.360332 10926 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0510 16:50:15.360344 10926 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0510 16:50:15.360357 10926 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0510 16:50:15.360368 10926 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0510 16:50:15.360380 10926 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0510 16:50:15.360391 10926 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.755682 | |
| I0510 16:50:15.360404 10926 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.159091 | |
| I0510 16:50:15.360419 10926 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.27324 (* 1 = 3.27324 loss) | |
| I0510 16:50:15.360432 10926 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 0.89003 (* 1 = 0.89003 loss) | |
| I0510 16:50:15.360447 10926 solver.cpp:245] Train net output #125: loss3/loss01 = 2.71458 (* 0.0909091 = 0.24678 loss) | |
| I0510 16:50:15.360461 10926 solver.cpp:245] Train net output #126: loss3/loss02 = 3.27969 (* 0.0909091 = 0.298153 loss) | |
| I0510 16:50:15.360476 10926 solver.cpp:245] Train net output #127: loss3/loss03 = 3.54972 (* 0.0909091 = 0.322702 loss) | |
| I0510 16:50:15.360491 10926 solver.cpp:245] Train net output #128: loss3/loss04 = 3.60447 (* 0.0909091 = 0.327679 loss) | |
| I0510 16:50:15.360504 10926 solver.cpp:245] Train net output #129: loss3/loss05 = 2.87926 (* 0.0909091 = 0.261751 loss) | |
| I0510 16:50:15.360519 10926 solver.cpp:245] Train net output #130: loss3/loss06 = 2.20435 (* 0.0909091 = 0.200395 loss) | |
| I0510 16:50:15.360532 10926 solver.cpp:245] Train net output #131: loss3/loss07 = 1.68825 (* 0.0909091 = 0.153478 loss) | |
| I0510 16:50:15.360548 10926 solver.cpp:245] Train net output #132: loss3/loss08 = 0.061687 (* 0.0909091 = 0.00560791 loss) | |
| I0510 16:50:15.360561 10926 solver.cpp:245] Train net output #133: loss3/loss09 = 0.00750574 (* 0.0909091 = 0.00068234 loss) | |
| I0510 16:50:15.360576 10926 solver.cpp:245] Train net output #134: loss3/loss10 = 0.00507204 (* 0.0909091 = 0.000461094 loss) | |
| I0510 16:50:15.360591 10926 solver.cpp:245] Train net output #135: loss3/loss11 = 0.00450401 (* 0.0909091 = 0.000409456 loss) | |
| I0510 16:50:15.360605 10926 solver.cpp:245] Train net output #136: loss3/loss12 = 0.00264659 (* 0.0909091 = 0.000240599 loss) | |
| I0510 16:50:15.360620 10926 solver.cpp:245] Train net output #137: loss3/loss13 = 0.00295465 (* 0.0909091 = 0.000268604 loss) | |
| I0510 16:50:15.360633 10926 solver.cpp:245] Train net output #138: loss3/loss14 = 0.00197411 (* 0.0909091 = 0.000179464 loss) | |
| I0510 16:50:15.360647 10926 solver.cpp:245] Train net output #139: loss3/loss15 = 0.00190961 (* 0.0909091 = 0.000173601 loss) | |
| I0510 16:50:15.360661 10926 solver.cpp:245] Train net output #140: loss3/loss16 = 0.00274239 (* 0.0909091 = 0.000249309 loss) | |
| I0510 16:50:15.360677 10926 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00181704 (* 0.0909091 = 0.000165186 loss) | |
| I0510 16:50:15.360690 10926 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00193498 (* 0.0909091 = 0.000175907 loss) | |
| I0510 16:50:15.360705 10926 solver.cpp:245] Train net output #143: loss3/loss19 = 0.0017052 (* 0.0909091 = 0.000155018 loss) | |
| I0510 16:50:15.360719 10926 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00150417 (* 0.0909091 = 0.000136743 loss) | |
| I0510 16:50:15.360734 10926 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00155232 (* 0.0909091 = 0.00014112 loss) | |
| I0510 16:50:15.360749 10926 solver.cpp:245] Train net output #146: loss3/loss22 = 0.000985388 (* 0.0909091 = 8.95807e-05 loss) | |
| I0510 16:50:15.360761 10926 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0510 16:50:15.360774 10926 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0510 16:50:15.360800 10926 solver.cpp:245] Train net output #149: total_confidence = 0.000241932 | |
| I0510 16:50:15.360813 10926 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.00035053 | |
| I0510 16:50:15.360826 10926 sgd_solver.cpp:106] Iteration 16500, lr = 0.001 | |
| I0510 16:52:43.008821 10926 solver.cpp:229] Iteration 17000, loss = 9.8274 | |
| I0510 16:52:43.008978 10926 solver.cpp:245] Train net output #0: loss1/accuracy = 0.104167 | |
| I0510 16:52:43.009001 10926 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.125 | |
| I0510 16:52:43.009014 10926 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0510 16:52:43.009027 10926 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0510 16:52:43.009040 10926 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.25 | |
| I0510 16:52:43.009053 10926 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.5 | |
| I0510 16:52:43.009066 10926 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.5 | |
| I0510 16:52:43.009079 10926 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.75 | |
| I0510 16:52:43.009093 10926 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.875 | |
| I0510 16:52:43.009105 10926 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0510 16:52:43.009130 10926 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.875 | |
| I0510 16:52:43.009147 10926 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0510 16:52:43.009160 10926 solver.cpp:245] Train net output #12: loss1/accuracy12 = 0.875 | |
| I0510 16:52:43.009173 10926 solver.cpp:245] Train net output #13: loss1/accuracy13 = 0.875 | |
| I0510 16:52:43.009186 10926 solver.cpp:245] Train net output #14: loss1/accuracy14 = 0.875 | |
| I0510 16:52:43.009198 10926 solver.cpp:245] Train net output #15: loss1/accuracy15 = 0.875 | |
| I0510 16:52:43.009212 10926 solver.cpp:245] Train net output #16: loss1/accuracy16 = 0.875 | |
| I0510 16:52:43.009223 10926 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0510 16:52:43.009237 10926 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0510 16:52:43.009248 10926 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0510 16:52:43.009259 10926 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0510 16:52:43.009271 10926 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0510 16:52:43.009284 10926 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0510 16:52:43.009295 10926 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.727273 | |
| I0510 16:52:43.009307 10926 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.25 | |
| I0510 16:52:43.009325 10926 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.04039 (* 0.3 = 0.912117 loss) | |
| I0510 16:52:43.009340 10926 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.01642 (* 0.3 = 0.304927 loss) | |
| I0510 16:52:43.009354 10926 solver.cpp:245] Train net output #27: loss1/loss01 = 2.86038 (* 0.0272727 = 0.0780104 loss) | |
| I0510 16:52:43.009369 10926 solver.cpp:245] Train net output #28: loss1/loss02 = 3.41647 (* 0.0272727 = 0.0931765 loss) | |
| I0510 16:52:43.009383 10926 solver.cpp:245] Train net output #29: loss1/loss03 = 3.77467 (* 0.0272727 = 0.102946 loss) | |
| I0510 16:52:43.009398 10926 solver.cpp:245] Train net output #30: loss1/loss04 = 2.6028 (* 0.0272727 = 0.0709853 loss) | |
| I0510 16:52:43.009413 10926 solver.cpp:245] Train net output #31: loss1/loss05 = 2.32201 (* 0.0272727 = 0.0633276 loss) | |
| I0510 16:52:43.009428 10926 solver.cpp:245] Train net output #32: loss1/loss06 = 1.97628 (* 0.0272727 = 0.0538986 loss) | |
| I0510 16:52:43.009441 10926 solver.cpp:245] Train net output #33: loss1/loss07 = 1.07756 (* 0.0272727 = 0.0293881 loss) | |
| I0510 16:52:43.009455 10926 solver.cpp:245] Train net output #34: loss1/loss08 = 0.466456 (* 0.0272727 = 0.0127215 loss) | |
| I0510 16:52:43.009471 10926 solver.cpp:245] Train net output #35: loss1/loss09 = 0.264761 (* 0.0272727 = 0.00722074 loss) | |
| I0510 16:52:43.009486 10926 solver.cpp:245] Train net output #36: loss1/loss10 = 0.492969 (* 0.0272727 = 0.0134446 loss) | |
| I0510 16:52:43.009501 10926 solver.cpp:245] Train net output #37: loss1/loss11 = 0.366626 (* 0.0272727 = 0.0099989 loss) | |
| I0510 16:52:43.009516 10926 solver.cpp:245] Train net output #38: loss1/loss12 = 0.470849 (* 0.0272727 = 0.0128413 loss) | |
| I0510 16:52:43.009531 10926 solver.cpp:245] Train net output #39: loss1/loss13 = 0.456905 (* 0.0272727 = 0.0124611 loss) | |
| I0510 16:52:43.009565 10926 solver.cpp:245] Train net output #40: loss1/loss14 = 0.438196 (* 0.0272727 = 0.0119508 loss) | |
| I0510 16:52:43.009582 10926 solver.cpp:245] Train net output #41: loss1/loss15 = 0.569243 (* 0.0272727 = 0.0155248 loss) | |
| I0510 16:52:43.009596 10926 solver.cpp:245] Train net output #42: loss1/loss16 = 0.51818 (* 0.0272727 = 0.0141322 loss) | |
| I0510 16:52:43.009611 10926 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0601639 (* 0.0272727 = 0.00164083 loss) | |
| I0510 16:52:43.009626 10926 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0828922 (* 0.0272727 = 0.0022607 loss) | |
| I0510 16:52:43.009640 10926 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0523668 (* 0.0272727 = 0.00142818 loss) | |
| I0510 16:52:43.009655 10926 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0228477 (* 0.0272727 = 0.000623118 loss) | |
| I0510 16:52:43.009670 10926 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0128843 (* 0.0272727 = 0.000351389 loss) | |
| I0510 16:52:43.009685 10926 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0342223 (* 0.0272727 = 0.000933336 loss) | |
| I0510 16:52:43.009698 10926 solver.cpp:245] Train net output #49: loss2/accuracy = 0.125 | |
| I0510 16:52:43.009711 10926 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.125 | |
| I0510 16:52:43.009723 10926 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0510 16:52:43.009735 10926 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0510 16:52:43.009747 10926 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.375 | |
| I0510 16:52:43.009760 10926 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.5 | |
| I0510 16:52:43.009773 10926 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.5 | |
| I0510 16:52:43.009784 10926 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.75 | |
| I0510 16:52:43.009796 10926 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.875 | |
| I0510 16:52:43.009809 10926 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.875 | |
| I0510 16:52:43.009821 10926 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.875 | |
| I0510 16:52:43.009835 10926 solver.cpp:245] Train net output #60: loss2/accuracy11 = 0.875 | |
| I0510 16:52:43.009846 10926 solver.cpp:245] Train net output #61: loss2/accuracy12 = 0.875 | |
| I0510 16:52:43.009858 10926 solver.cpp:245] Train net output #62: loss2/accuracy13 = 0.875 | |
| I0510 16:52:43.009871 10926 solver.cpp:245] Train net output #63: loss2/accuracy14 = 0.875 | |
| I0510 16:52:43.009886 10926 solver.cpp:245] Train net output #64: loss2/accuracy15 = 0.875 | |
| I0510 16:52:43.009899 10926 solver.cpp:245] Train net output #65: loss2/accuracy16 = 0.875 | |
| I0510 16:52:43.009912 10926 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0510 16:52:43.009923 10926 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0510 16:52:43.009935 10926 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0510 16:52:43.009948 10926 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0510 16:52:43.009959 10926 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0510 16:52:43.009971 10926 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0510 16:52:43.009982 10926 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.732955 | |
| I0510 16:52:43.009999 10926 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.270833 | |
| I0510 16:52:43.010013 10926 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 2.97641 (* 0.3 = 0.892924 loss) | |
| I0510 16:52:43.010028 10926 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.01376 (* 0.3 = 0.304128 loss) | |
| I0510 16:52:43.010042 10926 solver.cpp:245] Train net output #76: loss2/loss01 = 3.12391 (* 0.0272727 = 0.0851974 loss) | |
| I0510 16:52:43.010057 10926 solver.cpp:245] Train net output #77: loss2/loss02 = 3.64012 (* 0.0272727 = 0.0992759 loss) | |
| I0510 16:52:43.010083 10926 solver.cpp:245] Train net output #78: loss2/loss03 = 3.66652 (* 0.0272727 = 0.0999959 loss) | |
| I0510 16:52:43.010099 10926 solver.cpp:245] Train net output #79: loss2/loss04 = 2.15067 (* 0.0272727 = 0.0586548 loss) | |
| I0510 16:52:43.010113 10926 solver.cpp:245] Train net output #80: loss2/loss05 = 1.84608 (* 0.0272727 = 0.0503476 loss) | |
| I0510 16:52:43.010128 10926 solver.cpp:245] Train net output #81: loss2/loss06 = 2.20351 (* 0.0272727 = 0.0600957 loss) | |
| I0510 16:52:43.010143 10926 solver.cpp:245] Train net output #82: loss2/loss07 = 1.33035 (* 0.0272727 = 0.0362824 loss) | |
| I0510 16:52:43.010156 10926 solver.cpp:245] Train net output #83: loss2/loss08 = 0.427174 (* 0.0272727 = 0.0116502 loss) | |
| I0510 16:52:43.010170 10926 solver.cpp:245] Train net output #84: loss2/loss09 = 0.243419 (* 0.0272727 = 0.00663871 loss) | |
| I0510 16:52:43.010185 10926 solver.cpp:245] Train net output #85: loss2/loss10 = 0.37483 (* 0.0272727 = 0.0102226 loss) | |
| I0510 16:52:43.010200 10926 solver.cpp:245] Train net output #86: loss2/loss11 = 0.371349 (* 0.0272727 = 0.0101277 loss) | |
| I0510 16:52:43.010215 10926 solver.cpp:245] Train net output #87: loss2/loss12 = 0.456759 (* 0.0272727 = 0.0124571 loss) | |
| I0510 16:52:43.010228 10926 solver.cpp:245] Train net output #88: loss2/loss13 = 0.43991 (* 0.0272727 = 0.0119976 loss) | |
| I0510 16:52:43.010242 10926 solver.cpp:245] Train net output #89: loss2/loss14 = 0.533477 (* 0.0272727 = 0.0145494 loss) | |
| I0510 16:52:43.010257 10926 solver.cpp:245] Train net output #90: loss2/loss15 = 0.723883 (* 0.0272727 = 0.0197423 loss) | |
| I0510 16:52:43.010272 10926 solver.cpp:245] Train net output #91: loss2/loss16 = 0.664436 (* 0.0272727 = 0.018121 loss) | |
| I0510 16:52:43.010285 10926 solver.cpp:245] Train net output #92: loss2/loss17 = 0.00703548 (* 0.0272727 = 0.000191877 loss) | |
| I0510 16:52:43.010299 10926 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0132742 (* 0.0272727 = 0.000362024 loss) | |
| I0510 16:52:43.010314 10926 solver.cpp:245] Train net output #94: loss2/loss19 = 0.00433045 (* 0.0272727 = 0.000118103 loss) | |
| I0510 16:52:43.010329 10926 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00267868 (* 0.0272727 = 7.3055e-05 loss) | |
| I0510 16:52:43.010344 10926 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00146615 (* 0.0272727 = 3.9986e-05 loss) | |
| I0510 16:52:43.010359 10926 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0042466 (* 0.0272727 = 0.000115816 loss) | |
| I0510 16:52:43.010371 10926 solver.cpp:245] Train net output #98: loss3/accuracy = 0.145833 | |
| I0510 16:52:43.010385 10926 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.25 | |
| I0510 16:52:43.010396 10926 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0510 16:52:43.010408 10926 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0510 16:52:43.010421 10926 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.5 | |
| I0510 16:52:43.010433 10926 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.5 | |
| I0510 16:52:43.010445 10926 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.5 | |
| I0510 16:52:43.010457 10926 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.75 | |
| I0510 16:52:43.010469 10926 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.875 | |
| I0510 16:52:43.010480 10926 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0510 16:52:43.010493 10926 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.875 | |
| I0510 16:52:43.010505 10926 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0510 16:52:43.010517 10926 solver.cpp:245] Train net output #110: loss3/accuracy12 = 0.875 | |
| I0510 16:52:43.010529 10926 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0510 16:52:43.010542 10926 solver.cpp:245] Train net output #112: loss3/accuracy14 = 0.875 | |
| I0510 16:52:43.010557 10926 solver.cpp:245] Train net output #113: loss3/accuracy15 = 0.875 | |
| I0510 16:52:43.010570 10926 solver.cpp:245] Train net output #114: loss3/accuracy16 = 0.875 | |
| I0510 16:52:43.010591 10926 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0510 16:52:43.010606 10926 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0510 16:52:43.010617 10926 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0510 16:52:43.010629 10926 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0510 16:52:43.010640 10926 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0510 16:52:43.010653 10926 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0510 16:52:43.010664 10926 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.767045 | |
| I0510 16:52:43.010676 10926 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.333333 | |
| I0510 16:52:43.010690 10926 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 2.78689 (* 1 = 2.78689 loss) | |
| I0510 16:52:43.010704 10926 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 0.868357 (* 1 = 0.868357 loss) | |
| I0510 16:52:43.010718 10926 solver.cpp:245] Train net output #125: loss3/loss01 = 2.5044 (* 0.0909091 = 0.227673 loss) | |
| I0510 16:52:43.010732 10926 solver.cpp:245] Train net output #126: loss3/loss02 = 3.05408 (* 0.0909091 = 0.277644 loss) | |
| I0510 16:52:43.010746 10926 solver.cpp:245] Train net output #127: loss3/loss03 = 2.95835 (* 0.0909091 = 0.268941 loss) | |
| I0510 16:52:43.010761 10926 solver.cpp:245] Train net output #128: loss3/loss04 = 1.97323 (* 0.0909091 = 0.179385 loss) | |
| I0510 16:52:43.010774 10926 solver.cpp:245] Train net output #129: loss3/loss05 = 1.90798 (* 0.0909091 = 0.173452 loss) | |
| I0510 16:52:43.010788 10926 solver.cpp:245] Train net output #130: loss3/loss06 = 2.07054 (* 0.0909091 = 0.188231 loss) | |
| I0510 16:52:43.010803 10926 solver.cpp:245] Train net output #131: loss3/loss07 = 1.14999 (* 0.0909091 = 0.104545 loss) | |
| I0510 16:52:43.010818 10926 solver.cpp:245] Train net output #132: loss3/loss08 = 0.432375 (* 0.0909091 = 0.0393069 loss) | |
| I0510 16:52:43.010831 10926 solver.cpp:245] Train net output #133: loss3/loss09 = 0.1363 (* 0.0909091 = 0.0123909 loss) | |
| I0510 16:52:43.010845 10926 solver.cpp:245] Train net output #134: loss3/loss10 = 0.340514 (* 0.0909091 = 0.0309558 loss) | |
| I0510 16:52:43.010859 10926 solver.cpp:245] Train net output #135: loss3/loss11 = 0.239888 (* 0.0909091 = 0.021808 loss) | |
| I0510 16:52:43.010874 10926 solver.cpp:245] Train net output #136: loss3/loss12 = 0.359719 (* 0.0909091 = 0.0327018 loss) | |
| I0510 16:52:43.010888 10926 solver.cpp:245] Train net output #137: loss3/loss13 = 0.264499 (* 0.0909091 = 0.0240454 loss) | |
| I0510 16:52:43.010902 10926 solver.cpp:245] Train net output #138: loss3/loss14 = 0.288676 (* 0.0909091 = 0.0262432 loss) | |
| I0510 16:52:43.010916 10926 solver.cpp:245] Train net output #139: loss3/loss15 = 0.447741 (* 0.0909091 = 0.0407037 loss) | |
| I0510 16:52:43.010933 10926 solver.cpp:245] Train net output #140: loss3/loss16 = 0.371219 (* 0.0909091 = 0.0337472 loss) | |
| I0510 16:52:43.010948 10926 solver.cpp:245] Train net output #141: loss3/loss17 = 0.0250585 (* 0.0909091 = 0.00227805 loss) | |
| I0510 16:52:43.010962 10926 solver.cpp:245] Train net output #142: loss3/loss18 = 0.0113028 (* 0.0909091 = 0.00102753 loss) | |
| I0510 16:52:43.010977 10926 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00700149 (* 0.0909091 = 0.000636499 loss) | |
| I0510 16:52:43.010992 10926 solver.cpp:245] Train net output #144: loss3/loss20 = 0.0104662 (* 0.0909091 = 0.000951469 loss) | |
| I0510 16:52:43.011005 10926 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00401428 (* 0.0909091 = 0.000364935 loss) | |
| I0510 16:52:43.011020 10926 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00267028 (* 0.0909091 = 0.000242753 loss) | |
| I0510 16:52:43.011032 10926 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0510 16:52:43.011047 10926 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0510 16:52:43.011059 10926 solver.cpp:245] Train net output #149: total_confidence = 0.000330611 | |
| I0510 16:52:43.011081 10926 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.00347521 | |
| I0510 16:52:43.011097 10926 sgd_solver.cpp:106] Iteration 17000, lr = 0.001 | |
| I0510 16:52:57.007583 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 40.3733 > 30) by scale factor 0.743066 | |
| I0510 16:53:23.880524 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 40.6214 > 30) by scale factor 0.738527 | |
| I0510 16:55:10.747875 10926 solver.cpp:229] Iteration 17500, loss = 9.96359 | |
| I0510 16:55:10.748000 10926 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0350877 | |
| I0510 16:55:10.748021 10926 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0510 16:55:10.748035 10926 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0510 16:55:10.748049 10926 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0510 16:55:10.748061 10926 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0 | |
| I0510 16:55:10.748075 10926 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.375 | |
| I0510 16:55:10.748087 10926 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.25 | |
| I0510 16:55:10.748100 10926 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.625 | |
| I0510 16:55:10.748113 10926 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.75 | |
| I0510 16:55:10.748126 10926 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.75 | |
| I0510 16:55:10.748139 10926 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.75 | |
| I0510 16:55:10.748152 10926 solver.cpp:245] Train net output #11: loss1/accuracy11 = 0.875 | |
| I0510 16:55:10.748164 10926 solver.cpp:245] Train net output #12: loss1/accuracy12 = 0.875 | |
| I0510 16:55:10.748178 10926 solver.cpp:245] Train net output #13: loss1/accuracy13 = 0.875 | |
| I0510 16:55:10.748189 10926 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0510 16:55:10.748201 10926 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0510 16:55:10.748214 10926 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0510 16:55:10.748225 10926 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0510 16:55:10.748237 10926 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0510 16:55:10.748248 10926 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0510 16:55:10.748260 10926 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0510 16:55:10.748273 10926 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0510 16:55:10.748286 10926 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0510 16:55:10.748297 10926 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.6875 | |
| I0510 16:55:10.748309 10926 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.245614 | |
| I0510 16:55:10.748327 10926 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.42683 (* 0.3 = 1.02805 loss) | |
| I0510 16:55:10.748342 10926 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.1523 (* 0.3 = 0.34569 loss) | |
| I0510 16:55:10.748356 10926 solver.cpp:245] Train net output #27: loss1/loss01 = 3.28877 (* 0.0272727 = 0.0896937 loss) | |
| I0510 16:55:10.748370 10926 solver.cpp:245] Train net output #28: loss1/loss02 = 3.02837 (* 0.0272727 = 0.0825919 loss) | |
| I0510 16:55:10.748385 10926 solver.cpp:245] Train net output #29: loss1/loss03 = 3.41442 (* 0.0272727 = 0.0931206 loss) | |
| I0510 16:55:10.748399 10926 solver.cpp:245] Train net output #30: loss1/loss04 = 4.20804 (* 0.0272727 = 0.114765 loss) | |
| I0510 16:55:10.748414 10926 solver.cpp:245] Train net output #31: loss1/loss05 = 3.12472 (* 0.0272727 = 0.0852197 loss) | |
| I0510 16:55:10.748428 10926 solver.cpp:245] Train net output #32: loss1/loss06 = 2.98367 (* 0.0272727 = 0.0813728 loss) | |
| I0510 16:55:10.748442 10926 solver.cpp:245] Train net output #33: loss1/loss07 = 1.39326 (* 0.0272727 = 0.0379981 loss) | |
| I0510 16:55:10.748456 10926 solver.cpp:245] Train net output #34: loss1/loss08 = 1.22007 (* 0.0272727 = 0.0332746 loss) | |
| I0510 16:55:10.748471 10926 solver.cpp:245] Train net output #35: loss1/loss09 = 1.12177 (* 0.0272727 = 0.0305938 loss) | |
| I0510 16:55:10.748486 10926 solver.cpp:245] Train net output #36: loss1/loss10 = 1.15856 (* 0.0272727 = 0.0315972 loss) | |
| I0510 16:55:10.748499 10926 solver.cpp:245] Train net output #37: loss1/loss11 = 0.651638 (* 0.0272727 = 0.0177719 loss) | |
| I0510 16:55:10.748514 10926 solver.cpp:245] Train net output #38: loss1/loss12 = 0.894293 (* 0.0272727 = 0.0243898 loss) | |
| I0510 16:55:10.748528 10926 solver.cpp:245] Train net output #39: loss1/loss13 = 0.606664 (* 0.0272727 = 0.0165454 loss) | |
| I0510 16:55:10.748563 10926 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0347929 (* 0.0272727 = 0.000948897 loss) | |
| I0510 16:55:10.748579 10926 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0409256 (* 0.0272727 = 0.00111615 loss) | |
| I0510 16:55:10.748594 10926 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0287536 (* 0.0272727 = 0.000784188 loss) | |
| I0510 16:55:10.748608 10926 solver.cpp:245] Train net output #43: loss1/loss17 = 0.00669975 (* 0.0272727 = 0.000182721 loss) | |
| I0510 16:55:10.748622 10926 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0041648 (* 0.0272727 = 0.000113585 loss) | |
| I0510 16:55:10.748637 10926 solver.cpp:245] Train net output #45: loss1/loss19 = 0.00355881 (* 0.0272727 = 9.70584e-05 loss) | |
| I0510 16:55:10.748652 10926 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00647928 (* 0.0272727 = 0.000176708 loss) | |
| I0510 16:55:10.748667 10926 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00485419 (* 0.0272727 = 0.000132387 loss) | |
| I0510 16:55:10.748680 10926 solver.cpp:245] Train net output #48: loss1/loss22 = 0.00392762 (* 0.0272727 = 0.000107117 loss) | |
| I0510 16:55:10.748693 10926 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0175439 | |
| I0510 16:55:10.748705 10926 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0510 16:55:10.748718 10926 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0510 16:55:10.748730 10926 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0510 16:55:10.748741 10926 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.125 | |
| I0510 16:55:10.748754 10926 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.125 | |
| I0510 16:55:10.748766 10926 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.25 | |
| I0510 16:55:10.748778 10926 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.625 | |
| I0510 16:55:10.748791 10926 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.75 | |
| I0510 16:55:10.748803 10926 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.75 | |
| I0510 16:55:10.748816 10926 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.75 | |
| I0510 16:55:10.748827 10926 solver.cpp:245] Train net output #60: loss2/accuracy11 = 0.875 | |
| I0510 16:55:10.748839 10926 solver.cpp:245] Train net output #61: loss2/accuracy12 = 0.875 | |
| I0510 16:55:10.748852 10926 solver.cpp:245] Train net output #62: loss2/accuracy13 = 0.875 | |
| I0510 16:55:10.748864 10926 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0510 16:55:10.748879 10926 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0510 16:55:10.748893 10926 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0510 16:55:10.748904 10926 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0510 16:55:10.748916 10926 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0510 16:55:10.748929 10926 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0510 16:55:10.748940 10926 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0510 16:55:10.748952 10926 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0510 16:55:10.748965 10926 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0510 16:55:10.748976 10926 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.681818 | |
| I0510 16:55:10.748988 10926 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.192982 | |
| I0510 16:55:10.749002 10926 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.3331 (* 0.3 = 0.99993 loss) | |
| I0510 16:55:10.749020 10926 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.14601 (* 0.3 = 0.343803 loss) | |
| I0510 16:55:10.749035 10926 solver.cpp:245] Train net output #76: loss2/loss01 = 3.48855 (* 0.0272727 = 0.0951423 loss) | |
| I0510 16:55:10.749049 10926 solver.cpp:245] Train net output #77: loss2/loss02 = 3.29007 (* 0.0272727 = 0.089729 loss) | |
| I0510 16:55:10.749075 10926 solver.cpp:245] Train net output #78: loss2/loss03 = 3.64298 (* 0.0272727 = 0.0993539 loss) | |
| I0510 16:55:10.749090 10926 solver.cpp:245] Train net output #79: loss2/loss04 = 3.3791 (* 0.0272727 = 0.0921572 loss) | |
| I0510 16:55:10.749105 10926 solver.cpp:245] Train net output #80: loss2/loss05 = 3.03281 (* 0.0272727 = 0.0827131 loss) | |
| I0510 16:55:10.749115 10926 solver.cpp:245] Train net output #81: loss2/loss06 = 2.93673 (* 0.0272727 = 0.0800926 loss) | |
| I0510 16:55:10.749147 10926 solver.cpp:245] Train net output #82: loss2/loss07 = 1.51941 (* 0.0272727 = 0.0414384 loss) | |
| I0510 16:55:10.749162 10926 solver.cpp:245] Train net output #83: loss2/loss08 = 1.13465 (* 0.0272727 = 0.030945 loss) | |
| I0510 16:55:10.749177 10926 solver.cpp:245] Train net output #84: loss2/loss09 = 0.916273 (* 0.0272727 = 0.0249893 loss) | |
| I0510 16:55:10.749192 10926 solver.cpp:245] Train net output #85: loss2/loss10 = 1.1462 (* 0.0272727 = 0.03126 loss) | |
| I0510 16:55:10.749207 10926 solver.cpp:245] Train net output #86: loss2/loss11 = 0.557564 (* 0.0272727 = 0.0152063 loss) | |
| I0510 16:55:10.749222 10926 solver.cpp:245] Train net output #87: loss2/loss12 = 0.618235 (* 0.0272727 = 0.016861 loss) | |
| I0510 16:55:10.749235 10926 solver.cpp:245] Train net output #88: loss2/loss13 = 0.617618 (* 0.0272727 = 0.0168441 loss) | |
| I0510 16:55:10.749249 10926 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0321949 (* 0.0272727 = 0.000878042 loss) | |
| I0510 16:55:10.749264 10926 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0242738 (* 0.0272727 = 0.000662014 loss) | |
| I0510 16:55:10.749279 10926 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0171471 (* 0.0272727 = 0.000467649 loss) | |
| I0510 16:55:10.749292 10926 solver.cpp:245] Train net output #92: loss2/loss17 = 0.00377815 (* 0.0272727 = 0.00010304 loss) | |
| I0510 16:55:10.749306 10926 solver.cpp:245] Train net output #93: loss2/loss18 = 0.00446407 (* 0.0272727 = 0.000121747 loss) | |
| I0510 16:55:10.749321 10926 solver.cpp:245] Train net output #94: loss2/loss19 = 0.00143734 (* 0.0272727 = 3.92001e-05 loss) | |
| I0510 16:55:10.749336 10926 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00252431 (* 0.0272727 = 6.88449e-05 loss) | |
| I0510 16:55:10.749351 10926 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00209112 (* 0.0272727 = 5.70306e-05 loss) | |
| I0510 16:55:10.749364 10926 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00221299 (* 0.0272727 = 6.03542e-05 loss) | |
| I0510 16:55:10.749377 10926 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0877193 | |
| I0510 16:55:10.749389 10926 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0510 16:55:10.749402 10926 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0510 16:55:10.749413 10926 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0.25 | |
| I0510 16:55:10.749425 10926 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0510 16:55:10.749438 10926 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.375 | |
| I0510 16:55:10.749449 10926 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.25 | |
| I0510 16:55:10.749461 10926 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.625 | |
| I0510 16:55:10.749474 10926 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.75 | |
| I0510 16:55:10.749485 10926 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.75 | |
| I0510 16:55:10.749496 10926 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.75 | |
| I0510 16:55:10.749508 10926 solver.cpp:245] Train net output #109: loss3/accuracy11 = 0.875 | |
| I0510 16:55:10.749521 10926 solver.cpp:245] Train net output #110: loss3/accuracy12 = 0.875 | |
| I0510 16:55:10.749532 10926 solver.cpp:245] Train net output #111: loss3/accuracy13 = 0.875 | |
| I0510 16:55:10.749544 10926 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0510 16:55:10.749555 10926 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0510 16:55:10.749568 10926 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0510 16:55:10.749591 10926 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0510 16:55:10.749605 10926 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0510 16:55:10.749616 10926 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0510 16:55:10.749629 10926 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0510 16:55:10.749640 10926 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0510 16:55:10.749652 10926 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0510 16:55:10.749663 10926 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.704545 | |
| I0510 16:55:10.749676 10926 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.22807 | |
| I0510 16:55:10.749691 10926 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.18071 (* 1 = 3.18071 loss) | |
| I0510 16:55:10.749704 10926 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.05787 (* 1 = 1.05787 loss) | |
| I0510 16:55:10.749719 10926 solver.cpp:245] Train net output #125: loss3/loss01 = 3.26276 (* 0.0909091 = 0.296615 loss) | |
| I0510 16:55:10.749733 10926 solver.cpp:245] Train net output #126: loss3/loss02 = 3.21689 (* 0.0909091 = 0.292444 loss) | |
| I0510 16:55:10.749747 10926 solver.cpp:245] Train net output #127: loss3/loss03 = 2.77944 (* 0.0909091 = 0.252677 loss) | |
| I0510 16:55:10.749761 10926 solver.cpp:245] Train net output #128: loss3/loss04 = 3.57282 (* 0.0909091 = 0.324802 loss) | |
| I0510 16:55:10.749775 10926 solver.cpp:245] Train net output #129: loss3/loss05 = 2.55672 (* 0.0909091 = 0.232429 loss) | |
| I0510 16:55:10.749790 10926 solver.cpp:245] Train net output #130: loss3/loss06 = 2.74539 (* 0.0909091 = 0.249581 loss) | |
| I0510 16:55:10.749804 10926 solver.cpp:245] Train net output #131: loss3/loss07 = 1.32082 (* 0.0909091 = 0.120075 loss) | |
| I0510 16:55:10.749819 10926 solver.cpp:245] Train net output #132: loss3/loss08 = 1.12887 (* 0.0909091 = 0.102624 loss) | |
| I0510 16:55:10.749832 10926 solver.cpp:245] Train net output #133: loss3/loss09 = 1.03945 (* 0.0909091 = 0.0944958 loss) | |
| I0510 16:55:10.749846 10926 solver.cpp:245] Train net output #134: loss3/loss10 = 1.16159 (* 0.0909091 = 0.105599 loss) | |
| I0510 16:55:10.749861 10926 solver.cpp:245] Train net output #135: loss3/loss11 = 0.58985 (* 0.0909091 = 0.0536227 loss) | |
| I0510 16:55:10.749874 10926 solver.cpp:245] Train net output #136: loss3/loss12 = 0.725405 (* 0.0909091 = 0.0659459 loss) | |
| I0510 16:55:10.749888 10926 solver.cpp:245] Train net output #137: loss3/loss13 = 0.691383 (* 0.0909091 = 0.062853 loss) | |
| I0510 16:55:10.749903 10926 solver.cpp:245] Train net output #138: loss3/loss14 = 0.00853372 (* 0.0909091 = 0.000775793 loss) | |
| I0510 16:55:10.749917 10926 solver.cpp:245] Train net output #139: loss3/loss15 = 0.00642432 (* 0.0909091 = 0.000584029 loss) | |
| I0510 16:55:10.749934 10926 solver.cpp:245] Train net output #140: loss3/loss16 = 0.00489066 (* 0.0909091 = 0.000444605 loss) | |
| I0510 16:55:10.749949 10926 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00387225 (* 0.0909091 = 0.000352023 loss) | |
| I0510 16:55:10.749964 10926 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00240899 (* 0.0909091 = 0.000218999 loss) | |
| I0510 16:55:10.749979 10926 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00178158 (* 0.0909091 = 0.000161962 loss) | |
| I0510 16:55:10.749994 10926 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00133208 (* 0.0909091 = 0.000121098 loss) | |
| I0510 16:55:10.750008 10926 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00104593 (* 0.0909091 = 9.50844e-05 loss) | |
| I0510 16:55:10.750022 10926 solver.cpp:245] Train net output #146: loss3/loss22 = 0.000788536 (* 0.0909091 = 7.1685e-05 loss) | |
| I0510 16:55:10.750036 10926 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0510 16:55:10.750047 10926 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0510 16:55:10.750058 10926 solver.cpp:245] Train net output #149: total_confidence = 1.20017e-05 | |
| I0510 16:55:10.750083 10926 solver.cpp:245] Train net output #150: total_confidence_not_rec = 5.32469e-05 | |
| I0510 16:55:10.750099 10926 sgd_solver.cpp:106] Iteration 17500, lr = 0.001 | |
| I0510 16:57:07.705421 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 35.8446 > 30) by scale factor 0.836947 | |
| I0510 16:57:37.947348 10926 solver.cpp:229] Iteration 18000, loss = 9.84276 | |
| I0510 16:57:37.947496 10926 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0869565 | |
| I0510 16:57:37.947518 10926 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.25 | |
| I0510 16:57:37.947533 10926 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.125 | |
| I0510 16:57:37.947546 10926 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0510 16:57:37.947559 10926 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.375 | |
| I0510 16:57:37.947572 10926 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.375 | |
| I0510 16:57:37.947585 10926 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.625 | |
| I0510 16:57:37.947598 10926 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.625 | |
| I0510 16:57:37.947610 10926 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.625 | |
| I0510 16:57:37.947623 10926 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0510 16:57:37.947635 10926 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0510 16:57:37.947649 10926 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0510 16:57:37.947660 10926 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0510 16:57:37.947672 10926 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0510 16:57:37.947685 10926 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0510 16:57:37.947696 10926 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0510 16:57:37.947708 10926 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0510 16:57:37.947721 10926 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0510 16:57:37.947732 10926 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0510 16:57:37.947744 10926 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0510 16:57:37.947757 10926 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0510 16:57:37.947768 10926 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0510 16:57:37.947780 10926 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0510 16:57:37.947791 10926 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.75 | |
| I0510 16:57:37.947804 10926 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.282609 | |
| I0510 16:57:37.947820 10926 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 2.9654 (* 0.3 = 0.88962 loss) | |
| I0510 16:57:37.947835 10926 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 0.90656 (* 0.3 = 0.271968 loss) | |
| I0510 16:57:37.947850 10926 solver.cpp:245] Train net output #27: loss1/loss01 = 3.1633 (* 0.0272727 = 0.0862717 loss) | |
| I0510 16:57:37.947865 10926 solver.cpp:245] Train net output #28: loss1/loss02 = 2.81516 (* 0.0272727 = 0.0767771 loss) | |
| I0510 16:57:37.947882 10926 solver.cpp:245] Train net output #29: loss1/loss03 = 3.10229 (* 0.0272727 = 0.0846079 loss) | |
| I0510 16:57:37.947896 10926 solver.cpp:245] Train net output #30: loss1/loss04 = 2.80812 (* 0.0272727 = 0.0765852 loss) | |
| I0510 16:57:37.947911 10926 solver.cpp:245] Train net output #31: loss1/loss05 = 1.85129 (* 0.0272727 = 0.0504897 loss) | |
| I0510 16:57:37.947926 10926 solver.cpp:245] Train net output #32: loss1/loss06 = 1.58109 (* 0.0272727 = 0.0431207 loss) | |
| I0510 16:57:37.947939 10926 solver.cpp:245] Train net output #33: loss1/loss07 = 1.40415 (* 0.0272727 = 0.038295 loss) | |
| I0510 16:57:37.947953 10926 solver.cpp:245] Train net output #34: loss1/loss08 = 1.77503 (* 0.0272727 = 0.0484099 loss) | |
| I0510 16:57:37.947968 10926 solver.cpp:245] Train net output #35: loss1/loss09 = 0.0970006 (* 0.0272727 = 0.00264547 loss) | |
| I0510 16:57:37.947983 10926 solver.cpp:245] Train net output #36: loss1/loss10 = 0.0313281 (* 0.0272727 = 0.000854403 loss) | |
| I0510 16:57:37.947999 10926 solver.cpp:245] Train net output #37: loss1/loss11 = 0.0348202 (* 0.0272727 = 0.000949643 loss) | |
| I0510 16:57:37.948012 10926 solver.cpp:245] Train net output #38: loss1/loss12 = 0.00911186 (* 0.0272727 = 0.000248505 loss) | |
| I0510 16:57:37.948027 10926 solver.cpp:245] Train net output #39: loss1/loss13 = 0.013521 (* 0.0272727 = 0.000368753 loss) | |
| I0510 16:57:37.948060 10926 solver.cpp:245] Train net output #40: loss1/loss14 = 0.00771452 (* 0.0272727 = 0.000210396 loss) | |
| I0510 16:57:37.948076 10926 solver.cpp:245] Train net output #41: loss1/loss15 = 0.00832369 (* 0.0272727 = 0.00022701 loss) | |
| I0510 16:57:37.948091 10926 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0070119 (* 0.0272727 = 0.000191234 loss) | |
| I0510 16:57:37.948107 10926 solver.cpp:245] Train net output #43: loss1/loss17 = 0.00210171 (* 0.0272727 = 5.73193e-05 loss) | |
| I0510 16:57:37.948120 10926 solver.cpp:245] Train net output #44: loss1/loss18 = 0.00119401 (* 0.0272727 = 3.25638e-05 loss) | |
| I0510 16:57:37.948135 10926 solver.cpp:245] Train net output #45: loss1/loss19 = 0.000993136 (* 0.0272727 = 2.70855e-05 loss) | |
| I0510 16:57:37.948149 10926 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0027932 (* 0.0272727 = 7.61781e-05 loss) | |
| I0510 16:57:37.948164 10926 solver.cpp:245] Train net output #47: loss1/loss21 = 0.000764387 (* 0.0272727 = 2.08469e-05 loss) | |
| I0510 16:57:37.948179 10926 solver.cpp:245] Train net output #48: loss1/loss22 = 0.00150301 (* 0.0272727 = 4.0991e-05 loss) | |
| I0510 16:57:37.948191 10926 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0869565 | |
| I0510 16:57:37.948204 10926 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.25 | |
| I0510 16:57:37.948216 10926 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0.25 | |
| I0510 16:57:37.948228 10926 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0510 16:57:37.948240 10926 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.125 | |
| I0510 16:57:37.948251 10926 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.5 | |
| I0510 16:57:37.948263 10926 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.5 | |
| I0510 16:57:37.948276 10926 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.625 | |
| I0510 16:57:37.948287 10926 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.625 | |
| I0510 16:57:37.948299 10926 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0510 16:57:37.948312 10926 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0510 16:57:37.948323 10926 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0510 16:57:37.948335 10926 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0510 16:57:37.948348 10926 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0510 16:57:37.948362 10926 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0510 16:57:37.948374 10926 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0510 16:57:37.948386 10926 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0510 16:57:37.948398 10926 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0510 16:57:37.948410 10926 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0510 16:57:37.948421 10926 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0510 16:57:37.948433 10926 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0510 16:57:37.948446 10926 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0510 16:57:37.948457 10926 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0510 16:57:37.948468 10926 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.755682 | |
| I0510 16:57:37.948482 10926 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.23913 | |
| I0510 16:57:37.948495 10926 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 2.97119 (* 0.3 = 0.891358 loss) | |
| I0510 16:57:37.948513 10926 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 0.899425 (* 0.3 = 0.269828 loss) | |
| I0510 16:57:37.948529 10926 solver.cpp:245] Train net output #76: loss2/loss01 = 2.82486 (* 0.0272727 = 0.0770415 loss) | |
| I0510 16:57:37.948542 10926 solver.cpp:245] Train net output #77: loss2/loss02 = 2.79629 (* 0.0272727 = 0.0762623 loss) | |
| I0510 16:57:37.948568 10926 solver.cpp:245] Train net output #78: loss2/loss03 = 3.17046 (* 0.0272727 = 0.0864672 loss) | |
| I0510 16:57:37.948583 10926 solver.cpp:245] Train net output #79: loss2/loss04 = 3.44325 (* 0.0272727 = 0.0939068 loss) | |
| I0510 16:57:37.948598 10926 solver.cpp:245] Train net output #80: loss2/loss05 = 2.02344 (* 0.0272727 = 0.0551846 loss) | |
| I0510 16:57:37.948612 10926 solver.cpp:245] Train net output #81: loss2/loss06 = 1.56886 (* 0.0272727 = 0.042787 loss) | |
| I0510 16:57:37.948626 10926 solver.cpp:245] Train net output #82: loss2/loss07 = 1.41488 (* 0.0272727 = 0.0385876 loss) | |
| I0510 16:57:37.948640 10926 solver.cpp:245] Train net output #83: loss2/loss08 = 1.91732 (* 0.0272727 = 0.0522906 loss) | |
| I0510 16:57:37.948655 10926 solver.cpp:245] Train net output #84: loss2/loss09 = 0.103834 (* 0.0272727 = 0.00283182 loss) | |
| I0510 16:57:37.948670 10926 solver.cpp:245] Train net output #85: loss2/loss10 = 0.0766345 (* 0.0272727 = 0.00209003 loss) | |
| I0510 16:57:37.948684 10926 solver.cpp:245] Train net output #86: loss2/loss11 = 0.067525 (* 0.0272727 = 0.00184159 loss) | |
| I0510 16:57:37.948698 10926 solver.cpp:245] Train net output #87: loss2/loss12 = 0.0909257 (* 0.0272727 = 0.00247979 loss) | |
| I0510 16:57:37.948714 10926 solver.cpp:245] Train net output #88: loss2/loss13 = 0.0764184 (* 0.0272727 = 0.00208414 loss) | |
| I0510 16:57:37.948727 10926 solver.cpp:245] Train net output #89: loss2/loss14 = 0.019975 (* 0.0272727 = 0.000544772 loss) | |
| I0510 16:57:37.948741 10926 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0308028 (* 0.0272727 = 0.000840076 loss) | |
| I0510 16:57:37.948755 10926 solver.cpp:245] Train net output #91: loss2/loss16 = 0.026265 (* 0.0272727 = 0.000716317 loss) | |
| I0510 16:57:37.948770 10926 solver.cpp:245] Train net output #92: loss2/loss17 = 0.00485408 (* 0.0272727 = 0.000132384 loss) | |
| I0510 16:57:37.948784 10926 solver.cpp:245] Train net output #93: loss2/loss18 = 0.00299678 (* 0.0272727 = 8.17305e-05 loss) | |
| I0510 16:57:37.948798 10926 solver.cpp:245] Train net output #94: loss2/loss19 = 0.00419483 (* 0.0272727 = 0.000114404 loss) | |
| I0510 16:57:37.948812 10926 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00569563 (* 0.0272727 = 0.000155335 loss) | |
| I0510 16:57:37.948827 10926 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00347234 (* 0.0272727 = 9.47002e-05 loss) | |
| I0510 16:57:37.948842 10926 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00275041 (* 0.0272727 = 7.50113e-05 loss) | |
| I0510 16:57:37.948853 10926 solver.cpp:245] Train net output #98: loss3/accuracy = 0.108696 | |
| I0510 16:57:37.948866 10926 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0510 16:57:37.948879 10926 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.125 | |
| I0510 16:57:37.948890 10926 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0510 16:57:37.948904 10926 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0510 16:57:37.948915 10926 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.375 | |
| I0510 16:57:37.948930 10926 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.625 | |
| I0510 16:57:37.948942 10926 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.625 | |
| I0510 16:57:37.948954 10926 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.625 | |
| I0510 16:57:37.948966 10926 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0510 16:57:37.948978 10926 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0510 16:57:37.948990 10926 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0510 16:57:37.949002 10926 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0510 16:57:37.949013 10926 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0510 16:57:37.949025 10926 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0510 16:57:37.949038 10926 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0510 16:57:37.949048 10926 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0510 16:57:37.949071 10926 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0510 16:57:37.949084 10926 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0510 16:57:37.949096 10926 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0510 16:57:37.949108 10926 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0510 16:57:37.949137 10926 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0510 16:57:37.949152 10926 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0510 16:57:37.949163 10926 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.755682 | |
| I0510 16:57:37.949177 10926 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.23913 | |
| I0510 16:57:37.949190 10926 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 2.9131 (* 1 = 2.9131 loss) | |
| I0510 16:57:37.949204 10926 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 0.873938 (* 1 = 0.873938 loss) | |
| I0510 16:57:37.949218 10926 solver.cpp:245] Train net output #125: loss3/loss01 = 3.057 (* 0.0909091 = 0.277909 loss) | |
| I0510 16:57:37.949234 10926 solver.cpp:245] Train net output #126: loss3/loss02 = 2.73167 (* 0.0909091 = 0.248333 loss) | |
| I0510 16:57:37.949247 10926 solver.cpp:245] Train net output #127: loss3/loss03 = 2.76176 (* 0.0909091 = 0.251069 loss) | |
| I0510 16:57:37.949261 10926 solver.cpp:245] Train net output #128: loss3/loss04 = 2.95277 (* 0.0909091 = 0.268434 loss) | |
| I0510 16:57:37.949275 10926 solver.cpp:245] Train net output #129: loss3/loss05 = 2.02 (* 0.0909091 = 0.183636 loss) | |
| I0510 16:57:37.949290 10926 solver.cpp:245] Train net output #130: loss3/loss06 = 1.39279 (* 0.0909091 = 0.126617 loss) | |
| I0510 16:57:37.949304 10926 solver.cpp:245] Train net output #131: loss3/loss07 = 1.25264 (* 0.0909091 = 0.113876 loss) | |
| I0510 16:57:37.949318 10926 solver.cpp:245] Train net output #132: loss3/loss08 = 1.42251 (* 0.0909091 = 0.129319 loss) | |
| I0510 16:57:37.949332 10926 solver.cpp:245] Train net output #133: loss3/loss09 = 0.0956988 (* 0.0909091 = 0.00869989 loss) | |
| I0510 16:57:37.949347 10926 solver.cpp:245] Train net output #134: loss3/loss10 = 0.0609376 (* 0.0909091 = 0.00553978 loss) | |
| I0510 16:57:37.949362 10926 solver.cpp:245] Train net output #135: loss3/loss11 = 0.0400529 (* 0.0909091 = 0.00364118 loss) | |
| I0510 16:57:37.949376 10926 solver.cpp:245] Train net output #136: loss3/loss12 = 0.0182177 (* 0.0909091 = 0.00165615 loss) | |
| I0510 16:57:37.949391 10926 solver.cpp:245] Train net output #137: loss3/loss13 = 0.0100066 (* 0.0909091 = 0.000909689 loss) | |
| I0510 16:57:37.949405 10926 solver.cpp:245] Train net output #138: loss3/loss14 = 0.0065324 (* 0.0909091 = 0.000593855 loss) | |
| I0510 16:57:37.949419 10926 solver.cpp:245] Train net output #139: loss3/loss15 = 0.00304805 (* 0.0909091 = 0.000277095 loss) | |
| I0510 16:57:37.949434 10926 solver.cpp:245] Train net output #140: loss3/loss16 = 0.00229434 (* 0.0909091 = 0.000208577 loss) | |
| I0510 16:57:37.949447 10926 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00182058 (* 0.0909091 = 0.000165508 loss) | |
| I0510 16:57:37.949461 10926 solver.cpp:245] Train net output #142: loss3/loss18 = 0.000948579 (* 0.0909091 = 8.62344e-05 loss) | |
| I0510 16:57:37.949476 10926 solver.cpp:245] Train net output #143: loss3/loss19 = 0.000758165 (* 0.0909091 = 6.89241e-05 loss) | |
| I0510 16:57:37.949491 10926 solver.cpp:245] Train net output #144: loss3/loss20 = 0.000645898 (* 0.0909091 = 5.8718e-05 loss) | |
| I0510 16:57:37.949504 10926 solver.cpp:245] Train net output #145: loss3/loss21 = 0.000818657 (* 0.0909091 = 7.44234e-05 loss) | |
| I0510 16:57:37.949519 10926 solver.cpp:245] Train net output #146: loss3/loss22 = 0.000308682 (* 0.0909091 = 2.8062e-05 loss) | |
| I0510 16:57:37.949532 10926 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0510 16:57:37.949543 10926 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0510 16:57:37.949558 10926 solver.cpp:245] Train net output #149: total_confidence = 4.79492e-05 | |
| I0510 16:57:37.949584 10926 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.000309516 | |
| I0510 16:57:37.949599 10926 sgd_solver.cpp:106] Iteration 18000, lr = 0.001 | |
| I0510 16:58:00.727890 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 32.2691 > 30) by scale factor 0.929683 | |
| I0510 16:58:27.188683 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 36.35 > 30) by scale factor 0.82531 | |
| I0510 16:59:13.907856 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 32.2151 > 30) by scale factor 0.931239 | |
| I0510 17:00:05.364547 10926 solver.cpp:229] Iteration 18500, loss = 9.8786 | |
| I0510 17:00:05.364703 10926 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0666667 | |
| I0510 17:00:05.364724 10926 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0510 17:00:05.364738 10926 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0510 17:00:05.364754 10926 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.125 | |
| I0510 17:00:05.364768 10926 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.125 | |
| I0510 17:00:05.364780 10926 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.375 | |
| I0510 17:00:05.364794 10926 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.375 | |
| I0510 17:00:05.364806 10926 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.375 | |
| I0510 17:00:05.364820 10926 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.5 | |
| I0510 17:00:05.364832 10926 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.75 | |
| I0510 17:00:05.364845 10926 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.875 | |
| I0510 17:00:05.364858 10926 solver.cpp:245] Train net output #11: loss1/accuracy11 = 0.875 | |
| I0510 17:00:05.364871 10926 solver.cpp:245] Train net output #12: loss1/accuracy12 = 0.875 | |
| I0510 17:00:05.364883 10926 solver.cpp:245] Train net output #13: loss1/accuracy13 = 0.875 | |
| I0510 17:00:05.364895 10926 solver.cpp:245] Train net output #14: loss1/accuracy14 = 0.875 | |
| I0510 17:00:05.364917 10926 solver.cpp:245] Train net output #15: loss1/accuracy15 = 0.875 | |
| I0510 17:00:05.364928 10926 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0510 17:00:05.364940 10926 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0510 17:00:05.364953 10926 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0510 17:00:05.364965 10926 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0510 17:00:05.364977 10926 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0510 17:00:05.364989 10926 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0510 17:00:05.365002 10926 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0510 17:00:05.365015 10926 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.676136 | |
| I0510 17:00:05.365026 10926 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.233333 | |
| I0510 17:00:05.365052 10926 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.29983 (* 0.3 = 0.98995 loss) | |
| I0510 17:00:05.365067 10926 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.21963 (* 0.3 = 0.365889 loss) | |
| I0510 17:00:05.365082 10926 solver.cpp:245] Train net output #27: loss1/loss01 = 3.23366 (* 0.0272727 = 0.0881907 loss) | |
| I0510 17:00:05.365097 10926 solver.cpp:245] Train net output #28: loss1/loss02 = 3.68968 (* 0.0272727 = 0.100628 loss) | |
| I0510 17:00:05.365111 10926 solver.cpp:245] Train net output #29: loss1/loss03 = 3.7448 (* 0.0272727 = 0.102131 loss) | |
| I0510 17:00:05.365140 10926 solver.cpp:245] Train net output #30: loss1/loss04 = 3.06446 (* 0.0272727 = 0.0835762 loss) | |
| I0510 17:00:05.365157 10926 solver.cpp:245] Train net output #31: loss1/loss05 = 2.21545 (* 0.0272727 = 0.0604214 loss) | |
| I0510 17:00:05.365171 10926 solver.cpp:245] Train net output #32: loss1/loss06 = 2.24882 (* 0.0272727 = 0.0613314 loss) | |
| I0510 17:00:05.365186 10926 solver.cpp:245] Train net output #33: loss1/loss07 = 2.57234 (* 0.0272727 = 0.0701548 loss) | |
| I0510 17:00:05.365200 10926 solver.cpp:245] Train net output #34: loss1/loss08 = 2.41418 (* 0.0272727 = 0.0658414 loss) | |
| I0510 17:00:05.365214 10926 solver.cpp:245] Train net output #35: loss1/loss09 = 1.16942 (* 0.0272727 = 0.0318934 loss) | |
| I0510 17:00:05.365229 10926 solver.cpp:245] Train net output #36: loss1/loss10 = 0.380021 (* 0.0272727 = 0.0103642 loss) | |
| I0510 17:00:05.365243 10926 solver.cpp:245] Train net output #37: loss1/loss11 = 0.497893 (* 0.0272727 = 0.0135789 loss) | |
| I0510 17:00:05.365258 10926 solver.cpp:245] Train net output #38: loss1/loss12 = 0.633125 (* 0.0272727 = 0.0172671 loss) | |
| I0510 17:00:05.365295 10926 solver.cpp:245] Train net output #39: loss1/loss13 = 0.670558 (* 0.0272727 = 0.0182879 loss) | |
| I0510 17:00:05.365311 10926 solver.cpp:245] Train net output #40: loss1/loss14 = 0.649964 (* 0.0272727 = 0.0177263 loss) | |
| I0510 17:00:05.365329 10926 solver.cpp:245] Train net output #41: loss1/loss15 = 0.65368 (* 0.0272727 = 0.0178276 loss) | |
| I0510 17:00:05.365345 10926 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0401894 (* 0.0272727 = 0.00109608 loss) | |
| I0510 17:00:05.365360 10926 solver.cpp:245] Train net output #43: loss1/loss17 = 0.00747197 (* 0.0272727 = 0.000203781 loss) | |
| I0510 17:00:05.365375 10926 solver.cpp:245] Train net output #44: loss1/loss18 = 0.00717307 (* 0.0272727 = 0.000195629 loss) | |
| I0510 17:00:05.365389 10926 solver.cpp:245] Train net output #45: loss1/loss19 = 0.00662314 (* 0.0272727 = 0.000180631 loss) | |
| I0510 17:00:05.365403 10926 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00443804 (* 0.0272727 = 0.000121037 loss) | |
| I0510 17:00:05.365418 10926 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00474589 (* 0.0272727 = 0.000129433 loss) | |
| I0510 17:00:05.365432 10926 solver.cpp:245] Train net output #48: loss1/loss22 = 0.00345588 (* 0.0272727 = 9.42514e-05 loss) | |
| I0510 17:00:05.365445 10926 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0833333 | |
| I0510 17:00:05.365458 10926 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0510 17:00:05.365470 10926 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0510 17:00:05.365483 10926 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0.125 | |
| I0510 17:00:05.365494 10926 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.125 | |
| I0510 17:00:05.365506 10926 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.375 | |
| I0510 17:00:05.365520 10926 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.375 | |
| I0510 17:00:05.365531 10926 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.25 | |
| I0510 17:00:05.365543 10926 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.5 | |
| I0510 17:00:05.365556 10926 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.75 | |
| I0510 17:00:05.365567 10926 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.875 | |
| I0510 17:00:05.365581 10926 solver.cpp:245] Train net output #60: loss2/accuracy11 = 0.875 | |
| I0510 17:00:05.365592 10926 solver.cpp:245] Train net output #61: loss2/accuracy12 = 0.875 | |
| I0510 17:00:05.365604 10926 solver.cpp:245] Train net output #62: loss2/accuracy13 = 0.875 | |
| I0510 17:00:05.365617 10926 solver.cpp:245] Train net output #63: loss2/accuracy14 = 0.875 | |
| I0510 17:00:05.365628 10926 solver.cpp:245] Train net output #64: loss2/accuracy15 = 0.875 | |
| I0510 17:00:05.365641 10926 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0510 17:00:05.365653 10926 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0510 17:00:05.365664 10926 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0510 17:00:05.365676 10926 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0510 17:00:05.365689 10926 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0510 17:00:05.365700 10926 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0510 17:00:05.365712 10926 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0510 17:00:05.365726 10926 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.6875 | |
| I0510 17:00:05.365739 10926 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.25 | |
| I0510 17:00:05.365753 10926 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.24815 (* 0.3 = 0.974446 loss) | |
| I0510 17:00:05.365768 10926 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.20955 (* 0.3 = 0.362864 loss) | |
| I0510 17:00:05.365779 10926 solver.cpp:245] Train net output #76: loss2/loss01 = 3.26479 (* 0.0272727 = 0.0890398 loss) | |
| I0510 17:00:05.365793 10926 solver.cpp:245] Train net output #77: loss2/loss02 = 2.94619 (* 0.0272727 = 0.0803505 loss) | |
| I0510 17:00:05.365820 10926 solver.cpp:245] Train net output #78: loss2/loss03 = 3.18756 (* 0.0272727 = 0.0869335 loss) | |
| I0510 17:00:05.365835 10926 solver.cpp:245] Train net output #79: loss2/loss04 = 3.12418 (* 0.0272727 = 0.0852048 loss) | |
| I0510 17:00:05.365850 10926 solver.cpp:245] Train net output #80: loss2/loss05 = 2.4142 (* 0.0272727 = 0.0658417 loss) | |
| I0510 17:00:05.365864 10926 solver.cpp:245] Train net output #81: loss2/loss06 = 2.14485 (* 0.0272727 = 0.0584959 loss) | |
| I0510 17:00:05.365880 10926 solver.cpp:245] Train net output #82: loss2/loss07 = 2.55991 (* 0.0272727 = 0.0698157 loss) | |
| I0510 17:00:05.365893 10926 solver.cpp:245] Train net output #83: loss2/loss08 = 2.27349 (* 0.0272727 = 0.0620043 loss) | |
| I0510 17:00:05.365917 10926 solver.cpp:245] Train net output #84: loss2/loss09 = 1.24602 (* 0.0272727 = 0.0339823 loss) | |
| I0510 17:00:05.365932 10926 solver.cpp:245] Train net output #85: loss2/loss10 = 0.336299 (* 0.0272727 = 0.00917179 loss) | |
| I0510 17:00:05.365947 10926 solver.cpp:245] Train net output #86: loss2/loss11 = 0.540567 (* 0.0272727 = 0.0147427 loss) | |
| I0510 17:00:05.365962 10926 solver.cpp:245] Train net output #87: loss2/loss12 = 0.712572 (* 0.0272727 = 0.0194338 loss) | |
| I0510 17:00:05.365975 10926 solver.cpp:245] Train net output #88: loss2/loss13 = 0.644973 (* 0.0272727 = 0.0175902 loss) | |
| I0510 17:00:05.365993 10926 solver.cpp:245] Train net output #89: loss2/loss14 = 0.61782 (* 0.0272727 = 0.0168496 loss) | |
| I0510 17:00:05.366006 10926 solver.cpp:245] Train net output #90: loss2/loss15 = 0.67908 (* 0.0272727 = 0.0185204 loss) | |
| I0510 17:00:05.366021 10926 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0113753 (* 0.0272727 = 0.000310235 loss) | |
| I0510 17:00:05.366036 10926 solver.cpp:245] Train net output #92: loss2/loss17 = 0.00631039 (* 0.0272727 = 0.000172102 loss) | |
| I0510 17:00:05.366050 10926 solver.cpp:245] Train net output #93: loss2/loss18 = 0.00239477 (* 0.0272727 = 6.53118e-05 loss) | |
| I0510 17:00:05.366065 10926 solver.cpp:245] Train net output #94: loss2/loss19 = 0.00132705 (* 0.0272727 = 3.61924e-05 loss) | |
| I0510 17:00:05.366078 10926 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00100559 (* 0.0272727 = 2.74251e-05 loss) | |
| I0510 17:00:05.366101 10926 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00251396 (* 0.0272727 = 6.85626e-05 loss) | |
| I0510 17:00:05.366116 10926 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00332127 (* 0.0272727 = 9.05802e-05 loss) | |
| I0510 17:00:05.366128 10926 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0666667 | |
| I0510 17:00:05.366140 10926 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.125 | |
| I0510 17:00:05.366158 10926 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0510 17:00:05.366170 10926 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0510 17:00:05.366183 10926 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0510 17:00:05.366194 10926 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.5 | |
| I0510 17:00:05.366206 10926 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.375 | |
| I0510 17:00:05.366219 10926 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.375 | |
| I0510 17:00:05.366230 10926 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.5 | |
| I0510 17:00:05.366242 10926 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.75 | |
| I0510 17:00:05.366255 10926 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.875 | |
| I0510 17:00:05.366266 10926 solver.cpp:245] Train net output #109: loss3/accuracy11 = 0.875 | |
| I0510 17:00:05.366278 10926 solver.cpp:245] Train net output #110: loss3/accuracy12 = 0.875 | |
| I0510 17:00:05.366291 10926 solver.cpp:245] Train net output #111: loss3/accuracy13 = 0.875 | |
| I0510 17:00:05.366302 10926 solver.cpp:245] Train net output #112: loss3/accuracy14 = 0.875 | |
| I0510 17:00:05.366314 10926 solver.cpp:245] Train net output #113: loss3/accuracy15 = 0.875 | |
| I0510 17:00:05.366328 10926 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0510 17:00:05.366349 10926 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0510 17:00:05.366364 10926 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0510 17:00:05.366375 10926 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0510 17:00:05.366391 10926 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0510 17:00:05.366403 10926 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0510 17:00:05.366415 10926 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0510 17:00:05.366427 10926 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.676136 | |
| I0510 17:00:05.366441 10926 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.216667 | |
| I0510 17:00:05.366454 10926 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.04758 (* 1 = 3.04758 loss) | |
| I0510 17:00:05.366468 10926 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.10885 (* 1 = 1.10885 loss) | |
| I0510 17:00:05.366482 10926 solver.cpp:245] Train net output #125: loss3/loss01 = 2.88998 (* 0.0909091 = 0.262725 loss) | |
| I0510 17:00:05.366497 10926 solver.cpp:245] Train net output #126: loss3/loss02 = 3.09562 (* 0.0909091 = 0.28142 loss) | |
| I0510 17:00:05.366511 10926 solver.cpp:245] Train net output #127: loss3/loss03 = 3.11799 (* 0.0909091 = 0.283453 loss) | |
| I0510 17:00:05.366525 10926 solver.cpp:245] Train net output #128: loss3/loss04 = 2.81655 (* 0.0909091 = 0.25605 loss) | |
| I0510 17:00:05.366539 10926 solver.cpp:245] Train net output #129: loss3/loss05 = 2.28579 (* 0.0909091 = 0.207799 loss) | |
| I0510 17:00:05.366554 10926 solver.cpp:245] Train net output #130: loss3/loss06 = 2.16115 (* 0.0909091 = 0.196468 loss) | |
| I0510 17:00:05.366567 10926 solver.cpp:245] Train net output #131: loss3/loss07 = 2.44686 (* 0.0909091 = 0.222442 loss) | |
| I0510 17:00:05.366588 10926 solver.cpp:245] Train net output #132: loss3/loss08 = 2.17 (* 0.0909091 = 0.197273 loss) | |
| I0510 17:00:05.366602 10926 solver.cpp:245] Train net output #133: loss3/loss09 = 0.921131 (* 0.0909091 = 0.0837391 loss) | |
| I0510 17:00:05.366616 10926 solver.cpp:245] Train net output #134: loss3/loss10 = 0.411107 (* 0.0909091 = 0.0373733 loss) | |
| I0510 17:00:05.366631 10926 solver.cpp:245] Train net output #135: loss3/loss11 = 0.478832 (* 0.0909091 = 0.0435302 loss) | |
| I0510 17:00:05.366650 10926 solver.cpp:245] Train net output #136: loss3/loss12 = 0.520147 (* 0.0909091 = 0.0472861 loss) | |
| I0510 17:00:05.366664 10926 solver.cpp:245] Train net output #137: loss3/loss13 = 0.570974 (* 0.0909091 = 0.0519067 loss) | |
| I0510 17:00:05.366678 10926 solver.cpp:245] Train net output #138: loss3/loss14 = 0.516578 (* 0.0909091 = 0.0469616 loss) | |
| I0510 17:00:05.366693 10926 solver.cpp:245] Train net output #139: loss3/loss15 = 0.669324 (* 0.0909091 = 0.0608476 loss) | |
| I0510 17:00:05.366708 10926 solver.cpp:245] Train net output #140: loss3/loss16 = 0.00590695 (* 0.0909091 = 0.000536995 loss) | |
| I0510 17:00:05.366722 10926 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00336519 (* 0.0909091 = 0.000305926 loss) | |
| I0510 17:00:05.366736 10926 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00182737 (* 0.0909091 = 0.000166125 loss) | |
| I0510 17:00:05.366750 10926 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00150126 (* 0.0909091 = 0.000136479 loss) | |
| I0510 17:00:05.366765 10926 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00087736 (* 0.0909091 = 7.976e-05 loss) | |
| I0510 17:00:05.366783 10926 solver.cpp:245] Train net output #145: loss3/loss21 = 0.000734949 (* 0.0909091 = 6.68136e-05 loss) | |
| I0510 17:00:05.366798 10926 solver.cpp:245] Train net output #146: loss3/loss22 = 0.000414383 (* 0.0909091 = 3.76711e-05 loss) | |
| I0510 17:00:05.366811 10926 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0510 17:00:05.366823 10926 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0510 17:00:05.366835 10926 solver.cpp:245] Train net output #149: total_confidence = 0.000104263 | |
| I0510 17:00:05.366858 10926 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.00031914 | |
| I0510 17:00:05.366873 10926 sgd_solver.cpp:106] Iteration 18500, lr = 0.001 | |
| I0510 17:00:57.911931 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 32.6481 > 30) by scale factor 0.918891 | |
| I0510 17:02:29.307373 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 31.1307 > 30) by scale factor 0.96368 | |
| I0510 17:02:33.351441 10926 solver.cpp:229] Iteration 19000, loss = 9.71609 | |
| I0510 17:02:33.351522 10926 solver.cpp:245] Train net output #0: loss1/accuracy = 0.106383 | |
| I0510 17:02:33.351542 10926 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0510 17:02:33.351565 10926 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.125 | |
| I0510 17:02:33.351578 10926 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.125 | |
| I0510 17:02:33.351591 10926 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.125 | |
| I0510 17:02:33.351604 10926 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.25 | |
| I0510 17:02:33.351618 10926 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.25 | |
| I0510 17:02:33.351630 10926 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.625 | |
| I0510 17:02:33.351642 10926 solver.cpp:245] Train net output #8: loss1/accuracy08 = 1 | |
| I0510 17:02:33.351655 10926 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0510 17:02:33.351667 10926 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0510 17:02:33.351680 10926 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0510 17:02:33.351692 10926 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0510 17:02:33.351708 10926 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0510 17:02:33.351721 10926 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0510 17:02:33.351733 10926 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0510 17:02:33.351747 10926 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0510 17:02:33.351758 10926 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0510 17:02:33.351770 10926 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0510 17:02:33.351781 10926 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0510 17:02:33.351797 10926 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0510 17:02:33.351809 10926 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0510 17:02:33.351822 10926 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0510 17:02:33.351835 10926 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.755682 | |
| I0510 17:02:33.351846 10926 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.297872 | |
| I0510 17:02:33.351863 10926 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 2.90365 (* 0.3 = 0.871096 loss) | |
| I0510 17:02:33.351878 10926 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 0.89579 (* 0.3 = 0.268737 loss) | |
| I0510 17:02:33.351893 10926 solver.cpp:245] Train net output #27: loss1/loss01 = 2.82312 (* 0.0272727 = 0.0769943 loss) | |
| I0510 17:02:33.351907 10926 solver.cpp:245] Train net output #28: loss1/loss02 = 3.35496 (* 0.0272727 = 0.0914989 loss) | |
| I0510 17:02:33.351922 10926 solver.cpp:245] Train net output #29: loss1/loss03 = 2.97592 (* 0.0272727 = 0.0811614 loss) | |
| I0510 17:02:33.351936 10926 solver.cpp:245] Train net output #30: loss1/loss04 = 3.24187 (* 0.0272727 = 0.0884145 loss) | |
| I0510 17:02:33.351950 10926 solver.cpp:245] Train net output #31: loss1/loss05 = 2.74816 (* 0.0272727 = 0.0749499 loss) | |
| I0510 17:02:33.351964 10926 solver.cpp:245] Train net output #32: loss1/loss06 = 2.95455 (* 0.0272727 = 0.0805785 loss) | |
| I0510 17:02:33.351979 10926 solver.cpp:245] Train net output #33: loss1/loss07 = 2.25571 (* 0.0272727 = 0.0615193 loss) | |
| I0510 17:02:33.351994 10926 solver.cpp:245] Train net output #34: loss1/loss08 = 0.267131 (* 0.0272727 = 0.00728538 loss) | |
| I0510 17:02:33.352008 10926 solver.cpp:245] Train net output #35: loss1/loss09 = 0.036959 (* 0.0272727 = 0.00100797 loss) | |
| I0510 17:02:33.352023 10926 solver.cpp:245] Train net output #36: loss1/loss10 = 0.0253048 (* 0.0272727 = 0.000690132 loss) | |
| I0510 17:02:33.352037 10926 solver.cpp:245] Train net output #37: loss1/loss11 = 0.0187907 (* 0.0272727 = 0.000512474 loss) | |
| I0510 17:02:33.352052 10926 solver.cpp:245] Train net output #38: loss1/loss12 = 0.0119713 (* 0.0272727 = 0.00032649 loss) | |
| I0510 17:02:33.352097 10926 solver.cpp:245] Train net output #39: loss1/loss13 = 0.0334414 (* 0.0272727 = 0.000912038 loss) | |
| I0510 17:02:33.352114 10926 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0163622 (* 0.0272727 = 0.000446243 loss) | |
| I0510 17:02:33.352128 10926 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0136488 (* 0.0272727 = 0.000372241 loss) | |
| I0510 17:02:33.352144 10926 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0119315 (* 0.0272727 = 0.000325405 loss) | |
| I0510 17:02:33.352157 10926 solver.cpp:245] Train net output #43: loss1/loss17 = 0.00822517 (* 0.0272727 = 0.000224323 loss) | |
| I0510 17:02:33.352172 10926 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0240626 (* 0.0272727 = 0.000656253 loss) | |
| I0510 17:02:33.352186 10926 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0149047 (* 0.0272727 = 0.000406492 loss) | |
| I0510 17:02:33.352201 10926 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00647809 (* 0.0272727 = 0.000176675 loss) | |
| I0510 17:02:33.352216 10926 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00780709 (* 0.0272727 = 0.000212921 loss) | |
| I0510 17:02:33.352229 10926 solver.cpp:245] Train net output #48: loss1/loss22 = 0.00824022 (* 0.0272727 = 0.000224733 loss) | |
| I0510 17:02:33.352242 10926 solver.cpp:245] Train net output #49: loss2/accuracy = 0.148936 | |
| I0510 17:02:33.352255 10926 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.125 | |
| I0510 17:02:33.352267 10926 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0510 17:02:33.352279 10926 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0.125 | |
| I0510 17:02:33.352291 10926 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.125 | |
| I0510 17:02:33.352303 10926 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.125 | |
| I0510 17:02:33.352315 10926 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.25 | |
| I0510 17:02:33.352327 10926 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.625 | |
| I0510 17:02:33.352339 10926 solver.cpp:245] Train net output #57: loss2/accuracy08 = 1 | |
| I0510 17:02:33.352351 10926 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0510 17:02:33.352360 10926 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0510 17:02:33.352367 10926 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0510 17:02:33.352375 10926 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0510 17:02:33.352387 10926 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0510 17:02:33.352399 10926 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0510 17:02:33.352411 10926 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0510 17:02:33.352423 10926 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0510 17:02:33.352435 10926 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0510 17:02:33.352447 10926 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0510 17:02:33.352458 10926 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0510 17:02:33.352470 10926 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0510 17:02:33.352490 10926 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0510 17:02:33.352501 10926 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0510 17:02:33.352514 10926 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.761364 | |
| I0510 17:02:33.352525 10926 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.255319 | |
| I0510 17:02:33.352540 10926 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.05956 (* 0.3 = 0.917867 loss) | |
| I0510 17:02:33.352553 10926 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 0.962036 (* 0.3 = 0.288611 loss) | |
| I0510 17:02:33.352573 10926 solver.cpp:245] Train net output #76: loss2/loss01 = 2.70396 (* 0.0272727 = 0.0737444 loss) | |
| I0510 17:02:33.352587 10926 solver.cpp:245] Train net output #77: loss2/loss02 = 3.48595 (* 0.0272727 = 0.0950715 loss) | |
| I0510 17:02:33.352612 10926 solver.cpp:245] Train net output #78: loss2/loss03 = 3.23096 (* 0.0272727 = 0.088117 loss) | |
| I0510 17:02:33.352627 10926 solver.cpp:245] Train net output #79: loss2/loss04 = 2.73614 (* 0.0272727 = 0.0746219 loss) | |
| I0510 17:02:33.352643 10926 solver.cpp:245] Train net output #80: loss2/loss05 = 2.71578 (* 0.0272727 = 0.0740668 loss) | |
| I0510 17:02:33.352656 10926 solver.cpp:245] Train net output #81: loss2/loss06 = 3.34553 (* 0.0272727 = 0.0912418 loss) | |
| I0510 17:02:33.352670 10926 solver.cpp:245] Train net output #82: loss2/loss07 = 2.22454 (* 0.0272727 = 0.0606694 loss) | |
| I0510 17:02:33.352684 10926 solver.cpp:245] Train net output #83: loss2/loss08 = 0.269659 (* 0.0272727 = 0.00735434 loss) | |
| I0510 17:02:33.352699 10926 solver.cpp:245] Train net output #84: loss2/loss09 = 0.0702716 (* 0.0272727 = 0.0019165 loss) | |
| I0510 17:02:33.352715 10926 solver.cpp:245] Train net output #85: loss2/loss10 = 0.0227746 (* 0.0272727 = 0.000621126 loss) | |
| I0510 17:02:33.352728 10926 solver.cpp:245] Train net output #86: loss2/loss11 = 0.0201974 (* 0.0272727 = 0.000550837 loss) | |
| I0510 17:02:33.352742 10926 solver.cpp:245] Train net output #87: loss2/loss12 = 0.0152695 (* 0.0272727 = 0.000416441 loss) | |
| I0510 17:02:33.352761 10926 solver.cpp:245] Train net output #88: loss2/loss13 = 0.0137491 (* 0.0272727 = 0.000374975 loss) | |
| I0510 17:02:33.352774 10926 solver.cpp:245] Train net output #89: loss2/loss14 = 0.00963455 (* 0.0272727 = 0.00026276 loss) | |
| I0510 17:02:33.352788 10926 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0153537 (* 0.0272727 = 0.000418737 loss) | |
| I0510 17:02:33.352802 10926 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0097601 (* 0.0272727 = 0.000266185 loss) | |
| I0510 17:02:33.352816 10926 solver.cpp:245] Train net output #92: loss2/loss17 = 0.00517538 (* 0.0272727 = 0.000141147 loss) | |
| I0510 17:02:33.352830 10926 solver.cpp:245] Train net output #93: loss2/loss18 = 0.017934 (* 0.0272727 = 0.00048911 loss) | |
| I0510 17:02:33.352848 10926 solver.cpp:245] Train net output #94: loss2/loss19 = 0.013843 (* 0.0272727 = 0.000377537 loss) | |
| I0510 17:02:33.352861 10926 solver.cpp:245] Train net output #95: loss2/loss20 = 0.0076612 (* 0.0272727 = 0.000208942 loss) | |
| I0510 17:02:33.352876 10926 solver.cpp:245] Train net output #96: loss2/loss21 = 0.0195128 (* 0.0272727 = 0.000532167 loss) | |
| I0510 17:02:33.352890 10926 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00795413 (* 0.0272727 = 0.000216931 loss) | |
| I0510 17:02:33.352902 10926 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0212766 | |
| I0510 17:02:33.352915 10926 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.125 | |
| I0510 17:02:33.352926 10926 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0510 17:02:33.352938 10926 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0510 17:02:33.352951 10926 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0510 17:02:33.352962 10926 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.375 | |
| I0510 17:02:33.352974 10926 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.5 | |
| I0510 17:02:33.352985 10926 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.625 | |
| I0510 17:02:33.352998 10926 solver.cpp:245] Train net output #106: loss3/accuracy08 = 1 | |
| I0510 17:02:33.353008 10926 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0510 17:02:33.353020 10926 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0510 17:02:33.353031 10926 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0510 17:02:33.353044 10926 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0510 17:02:33.353055 10926 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0510 17:02:33.353066 10926 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0510 17:02:33.353077 10926 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0510 17:02:33.353099 10926 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0510 17:02:33.353112 10926 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0510 17:02:33.353138 10926 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0510 17:02:33.353152 10926 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0510 17:02:33.353164 10926 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0510 17:02:33.353175 10926 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0510 17:02:33.353188 10926 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0510 17:02:33.353199 10926 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.738636 | |
| I0510 17:02:33.353210 10926 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.170213 | |
| I0510 17:02:33.353224 10926 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 2.845 (* 1 = 2.845 loss) | |
| I0510 17:02:33.353238 10926 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 0.817935 (* 1 = 0.817935 loss) | |
| I0510 17:02:33.353253 10926 solver.cpp:245] Train net output #125: loss3/loss01 = 2.56277 (* 0.0909091 = 0.232979 loss) | |
| I0510 17:02:33.353267 10926 solver.cpp:245] Train net output #126: loss3/loss02 = 3.00153 (* 0.0909091 = 0.272866 loss) | |
| I0510 17:02:33.353281 10926 solver.cpp:245] Train net output #127: loss3/loss03 = 2.8048 (* 0.0909091 = 0.254982 loss) | |
| I0510 17:02:33.353296 10926 solver.cpp:245] Train net output #128: loss3/loss04 = 2.66856 (* 0.0909091 = 0.242596 loss) | |
| I0510 17:02:33.353310 10926 solver.cpp:245] Train net output #129: loss3/loss05 = 2.13383 (* 0.0909091 = 0.193984 loss) | |
| I0510 17:02:33.353324 10926 solver.cpp:245] Train net output #130: loss3/loss06 = 2.60198 (* 0.0909091 = 0.236544 loss) | |
| I0510 17:02:33.353338 10926 solver.cpp:245] Train net output #131: loss3/loss07 = 1.96899 (* 0.0909091 = 0.178999 loss) | |
| I0510 17:02:33.353351 10926 solver.cpp:245] Train net output #132: loss3/loss08 = 0.261295 (* 0.0909091 = 0.0237541 loss) | |
| I0510 17:02:33.353366 10926 solver.cpp:245] Train net output #133: loss3/loss09 = 0.0189555 (* 0.0909091 = 0.00172323 loss) | |
| I0510 17:02:33.353380 10926 solver.cpp:245] Train net output #134: loss3/loss10 = 0.0124775 (* 0.0909091 = 0.00113432 loss) | |
| I0510 17:02:33.353394 10926 solver.cpp:245] Train net output #135: loss3/loss11 = 0.00695086 (* 0.0909091 = 0.000631896 loss) | |
| I0510 17:02:33.353409 10926 solver.cpp:245] Train net output #136: loss3/loss12 = 0.00561659 (* 0.0909091 = 0.000510599 loss) | |
| I0510 17:02:33.353423 10926 solver.cpp:245] Train net output #137: loss3/loss13 = 0.00354253 (* 0.0909091 = 0.000322048 loss) | |
| I0510 17:02:33.353437 10926 solver.cpp:245] Train net output #138: loss3/loss14 = 0.00483399 (* 0.0909091 = 0.000439454 loss) | |
| I0510 17:02:33.353451 10926 solver.cpp:245] Train net output #139: loss3/loss15 = 0.00196472 (* 0.0909091 = 0.000178611 loss) | |
| I0510 17:02:33.353472 10926 solver.cpp:245] Train net output #140: loss3/loss16 = 0.0017334 (* 0.0909091 = 0.000157582 loss) | |
| I0510 17:02:33.353487 10926 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00367934 (* 0.0909091 = 0.000334486 loss) | |
| I0510 17:02:33.353502 10926 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00198191 (* 0.0909091 = 0.000180174 loss) | |
| I0510 17:02:33.353515 10926 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00231409 (* 0.0909091 = 0.000210372 loss) | |
| I0510 17:02:33.353529 10926 solver.cpp:245] Train net output #144: loss3/loss20 = 0.0025092 (* 0.0909091 = 0.000228109 loss) | |
| I0510 17:02:33.353543 10926 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00179557 (* 0.0909091 = 0.000163233 loss) | |
| I0510 17:02:33.353562 10926 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00166959 (* 0.0909091 = 0.000151781 loss) | |
| I0510 17:02:33.353574 10926 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0510 17:02:33.353585 10926 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0510 17:02:33.353608 10926 solver.cpp:245] Train net output #149: total_confidence = 3.00713e-05 | |
| I0510 17:02:33.353622 10926 solver.cpp:245] Train net output #150: total_confidence_not_rec = 9.87317e-05 | |
| I0510 17:02:33.353636 10926 sgd_solver.cpp:106] Iteration 19000, lr = 0.001 | |
| I0510 17:02:50.307327 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 42.2598 > 30) by scale factor 0.709895 | |
| I0510 17:04:10.547212 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 31.3686 > 30) by scale factor 0.956371 | |
| I0510 17:05:01.430094 10926 solver.cpp:229] Iteration 19500, loss = 9.712 | |
| I0510 17:05:01.430224 10926 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0816327 | |
| I0510 17:05:01.430246 10926 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0510 17:05:01.430260 10926 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.25 | |
| I0510 17:05:01.430274 10926 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.125 | |
| I0510 17:05:01.430286 10926 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.125 | |
| I0510 17:05:01.430299 10926 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.375 | |
| I0510 17:05:01.430312 10926 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.25 | |
| I0510 17:05:01.430325 10926 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.375 | |
| I0510 17:05:01.430338 10926 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.875 | |
| I0510 17:05:01.430351 10926 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0510 17:05:01.430364 10926 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0510 17:05:01.430377 10926 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0510 17:05:01.430389 10926 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0510 17:05:01.430402 10926 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0510 17:05:01.430423 10926 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0510 17:05:01.430441 10926 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0510 17:05:01.430454 10926 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0510 17:05:01.430466 10926 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0510 17:05:01.430479 10926 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0510 17:05:01.430490 10926 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0510 17:05:01.430502 10926 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0510 17:05:01.430515 10926 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0510 17:05:01.430526 10926 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0510 17:05:01.430538 10926 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.715909 | |
| I0510 17:05:01.430552 10926 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.326531 | |
| I0510 17:05:01.430567 10926 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 2.97885 (* 0.3 = 0.893656 loss) | |
| I0510 17:05:01.430583 10926 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.00144 (* 0.3 = 0.300432 loss) | |
| I0510 17:05:01.430598 10926 solver.cpp:245] Train net output #27: loss1/loss01 = 3.08444 (* 0.0272727 = 0.084121 loss) | |
| I0510 17:05:01.430611 10926 solver.cpp:245] Train net output #28: loss1/loss02 = 2.57276 (* 0.0272727 = 0.0701663 loss) | |
| I0510 17:05:01.430626 10926 solver.cpp:245] Train net output #29: loss1/loss03 = 2.87823 (* 0.0272727 = 0.0784971 loss) | |
| I0510 17:05:01.430640 10926 solver.cpp:245] Train net output #30: loss1/loss04 = 3.41456 (* 0.0272727 = 0.0931243 loss) | |
| I0510 17:05:01.430655 10926 solver.cpp:245] Train net output #31: loss1/loss05 = 2.70096 (* 0.0272727 = 0.0736624 loss) | |
| I0510 17:05:01.430677 10926 solver.cpp:245] Train net output #32: loss1/loss06 = 2.71466 (* 0.0272727 = 0.0740362 loss) | |
| I0510 17:05:01.430694 10926 solver.cpp:245] Train net output #33: loss1/loss07 = 2.82 (* 0.0272727 = 0.0769091 loss) | |
| I0510 17:05:01.430709 10926 solver.cpp:245] Train net output #34: loss1/loss08 = 0.886525 (* 0.0272727 = 0.0241779 loss) | |
| I0510 17:05:01.430724 10926 solver.cpp:245] Train net output #35: loss1/loss09 = 0.129768 (* 0.0272727 = 0.00353913 loss) | |
| I0510 17:05:01.430739 10926 solver.cpp:245] Train net output #36: loss1/loss10 = 0.0736569 (* 0.0272727 = 0.00200882 loss) | |
| I0510 17:05:01.430753 10926 solver.cpp:245] Train net output #37: loss1/loss11 = 0.0425493 (* 0.0272727 = 0.00116044 loss) | |
| I0510 17:05:01.430768 10926 solver.cpp:245] Train net output #38: loss1/loss12 = 0.0330928 (* 0.0272727 = 0.00090253 loss) | |
| I0510 17:05:01.430783 10926 solver.cpp:245] Train net output #39: loss1/loss13 = 0.0250182 (* 0.0272727 = 0.000682313 loss) | |
| I0510 17:05:01.430816 10926 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0216953 (* 0.0272727 = 0.00059169 loss) | |
| I0510 17:05:01.430833 10926 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0213341 (* 0.0272727 = 0.000581839 loss) | |
| I0510 17:05:01.430847 10926 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0088066 (* 0.0272727 = 0.00024018 loss) | |
| I0510 17:05:01.430862 10926 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0110487 (* 0.0272727 = 0.000301329 loss) | |
| I0510 17:05:01.430879 10926 solver.cpp:245] Train net output #44: loss1/loss18 = 0.00604644 (* 0.0272727 = 0.000164903 loss) | |
| I0510 17:05:01.430894 10926 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0065255 (* 0.0272727 = 0.000177968 loss) | |
| I0510 17:05:01.430909 10926 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00389466 (* 0.0272727 = 0.000106218 loss) | |
| I0510 17:05:01.430923 10926 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00452923 (* 0.0272727 = 0.000123524 loss) | |
| I0510 17:05:01.430938 10926 solver.cpp:245] Train net output #48: loss1/loss22 = 0.00432789 (* 0.0272727 = 0.000118033 loss) | |
| I0510 17:05:01.430950 10926 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0408163 | |
| I0510 17:05:01.430963 10926 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.25 | |
| I0510 17:05:01.430975 10926 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0510 17:05:01.430987 10926 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0510 17:05:01.430999 10926 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.25 | |
| I0510 17:05:01.431011 10926 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.25 | |
| I0510 17:05:01.431025 10926 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.375 | |
| I0510 17:05:01.431036 10926 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.375 | |
| I0510 17:05:01.431048 10926 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.875 | |
| I0510 17:05:01.431061 10926 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0510 17:05:01.431072 10926 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0510 17:05:01.431084 10926 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0510 17:05:01.431097 10926 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0510 17:05:01.431107 10926 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0510 17:05:01.431119 10926 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0510 17:05:01.431131 10926 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0510 17:05:01.431143 10926 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0510 17:05:01.431155 10926 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0510 17:05:01.431166 10926 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0510 17:05:01.431179 10926 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0510 17:05:01.431190 10926 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0510 17:05:01.431201 10926 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0510 17:05:01.431213 10926 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0510 17:05:01.431226 10926 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.710227 | |
| I0510 17:05:01.431241 10926 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.306122 | |
| I0510 17:05:01.431255 10926 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 2.89971 (* 0.3 = 0.869913 loss) | |
| I0510 17:05:01.431270 10926 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 0.972883 (* 0.3 = 0.291865 loss) | |
| I0510 17:05:01.431284 10926 solver.cpp:245] Train net output #76: loss2/loss01 = 2.79068 (* 0.0272727 = 0.0761093 loss) | |
| I0510 17:05:01.431299 10926 solver.cpp:245] Train net output #77: loss2/loss02 = 2.98541 (* 0.0272727 = 0.0814203 loss) | |
| I0510 17:05:01.431324 10926 solver.cpp:245] Train net output #78: loss2/loss03 = 3.45825 (* 0.0272727 = 0.094316 loss) | |
| I0510 17:05:01.431340 10926 solver.cpp:245] Train net output #79: loss2/loss04 = 3.12017 (* 0.0272727 = 0.0850955 loss) | |
| I0510 17:05:01.431355 10926 solver.cpp:245] Train net output #80: loss2/loss05 = 2.68304 (* 0.0272727 = 0.0731739 loss) | |
| I0510 17:05:01.431370 10926 solver.cpp:245] Train net output #81: loss2/loss06 = 2.32285 (* 0.0272727 = 0.0633506 loss) | |
| I0510 17:05:01.431385 10926 solver.cpp:245] Train net output #82: loss2/loss07 = 2.86999 (* 0.0272727 = 0.0782726 loss) | |
| I0510 17:05:01.431398 10926 solver.cpp:245] Train net output #83: loss2/loss08 = 0.78589 (* 0.0272727 = 0.0214334 loss) | |
| I0510 17:05:01.431412 10926 solver.cpp:245] Train net output #84: loss2/loss09 = 0.133807 (* 0.0272727 = 0.00364928 loss) | |
| I0510 17:05:01.431427 10926 solver.cpp:245] Train net output #85: loss2/loss10 = 0.0452026 (* 0.0272727 = 0.0012328 loss) | |
| I0510 17:05:01.431442 10926 solver.cpp:245] Train net output #86: loss2/loss11 = 0.0289407 (* 0.0272727 = 0.000789293 loss) | |
| I0510 17:05:01.431457 10926 solver.cpp:245] Train net output #87: loss2/loss12 = 0.0176386 (* 0.0272727 = 0.000481053 loss) | |
| I0510 17:05:01.431473 10926 solver.cpp:245] Train net output #88: loss2/loss13 = 0.0159666 (* 0.0272727 = 0.000435452 loss) | |
| I0510 17:05:01.431483 10926 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0325662 (* 0.0272727 = 0.00088817 loss) | |
| I0510 17:05:01.431499 10926 solver.cpp:245] Train net output #90: loss2/loss15 = 0.00980337 (* 0.0272727 = 0.000267365 loss) | |
| I0510 17:05:01.431522 10926 solver.cpp:245] Train net output #91: loss2/loss16 = 0.00912917 (* 0.0272727 = 0.000248977 loss) | |
| I0510 17:05:01.431540 10926 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0042313 (* 0.0272727 = 0.000115399 loss) | |
| I0510 17:05:01.431553 10926 solver.cpp:245] Train net output #93: loss2/loss18 = 0.00552362 (* 0.0272727 = 0.000150644 loss) | |
| I0510 17:05:01.431568 10926 solver.cpp:245] Train net output #94: loss2/loss19 = 0.00252615 (* 0.0272727 = 6.88949e-05 loss) | |
| I0510 17:05:01.431582 10926 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00418191 (* 0.0272727 = 0.000114052 loss) | |
| I0510 17:05:01.431597 10926 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00191051 (* 0.0272727 = 5.21048e-05 loss) | |
| I0510 17:05:01.431612 10926 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00074932 (* 0.0272727 = 2.0436e-05 loss) | |
| I0510 17:05:01.431624 10926 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0204082 | |
| I0510 17:05:01.431638 10926 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0510 17:05:01.431650 10926 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0510 17:05:01.431663 10926 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0.125 | |
| I0510 17:05:01.431674 10926 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0510 17:05:01.431686 10926 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.125 | |
| I0510 17:05:01.431699 10926 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.25 | |
| I0510 17:05:01.431710 10926 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.5 | |
| I0510 17:05:01.431723 10926 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.875 | |
| I0510 17:05:01.431735 10926 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0510 17:05:01.431746 10926 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0510 17:05:01.431758 10926 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0510 17:05:01.431771 10926 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0510 17:05:01.431782 10926 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0510 17:05:01.431794 10926 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0510 17:05:01.431807 10926 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0510 17:05:01.431818 10926 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0510 17:05:01.431840 10926 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0510 17:05:01.431854 10926 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0510 17:05:01.431866 10926 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0510 17:05:01.431879 10926 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0510 17:05:01.431890 10926 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0510 17:05:01.431902 10926 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0510 17:05:01.431913 10926 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.710227 | |
| I0510 17:05:01.431928 10926 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.265306 | |
| I0510 17:05:01.431943 10926 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 2.85215 (* 1 = 2.85215 loss) | |
| I0510 17:05:01.431957 10926 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 0.886364 (* 1 = 0.886364 loss) | |
| I0510 17:05:01.431972 10926 solver.cpp:245] Train net output #125: loss3/loss01 = 2.58577 (* 0.0909091 = 0.23507 loss) | |
| I0510 17:05:01.431987 10926 solver.cpp:245] Train net output #126: loss3/loss02 = 2.62527 (* 0.0909091 = 0.238661 loss) | |
| I0510 17:05:01.432000 10926 solver.cpp:245] Train net output #127: loss3/loss03 = 2.75399 (* 0.0909091 = 0.250362 loss) | |
| I0510 17:05:01.432014 10926 solver.cpp:245] Train net output #128: loss3/loss04 = 2.85547 (* 0.0909091 = 0.259589 loss) | |
| I0510 17:05:01.432029 10926 solver.cpp:245] Train net output #129: loss3/loss05 = 2.36913 (* 0.0909091 = 0.215375 loss) | |
| I0510 17:05:01.432042 10926 solver.cpp:245] Train net output #130: loss3/loss06 = 2.37241 (* 0.0909091 = 0.215674 loss) | |
| I0510 17:05:01.432056 10926 solver.cpp:245] Train net output #131: loss3/loss07 = 2.32053 (* 0.0909091 = 0.210957 loss) | |
| I0510 17:05:01.432070 10926 solver.cpp:245] Train net output #132: loss3/loss08 = 0.757879 (* 0.0909091 = 0.0688981 loss) | |
| I0510 17:05:01.432085 10926 solver.cpp:245] Train net output #133: loss3/loss09 = 0.213133 (* 0.0909091 = 0.0193758 loss) | |
| I0510 17:05:01.432099 10926 solver.cpp:245] Train net output #134: loss3/loss10 = 0.0546572 (* 0.0909091 = 0.00496884 loss) | |
| I0510 17:05:01.432113 10926 solver.cpp:245] Train net output #135: loss3/loss11 = 0.0155059 (* 0.0909091 = 0.00140963 loss) | |
| I0510 17:05:01.432127 10926 solver.cpp:245] Train net output #136: loss3/loss12 = 0.0145534 (* 0.0909091 = 0.00132304 loss) | |
| I0510 17:05:01.432142 10926 solver.cpp:245] Train net output #137: loss3/loss13 = 0.00857132 (* 0.0909091 = 0.000779211 loss) | |
| I0510 17:05:01.432157 10926 solver.cpp:245] Train net output #138: loss3/loss14 = 0.00589898 (* 0.0909091 = 0.000536271 loss) | |
| I0510 17:05:01.432170 10926 solver.cpp:245] Train net output #139: loss3/loss15 = 0.00441754 (* 0.0909091 = 0.000401595 loss) | |
| I0510 17:05:01.432184 10926 solver.cpp:245] Train net output #140: loss3/loss16 = 0.00441769 (* 0.0909091 = 0.000401609 loss) | |
| I0510 17:05:01.432199 10926 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00243486 (* 0.0909091 = 0.000221351 loss) | |
| I0510 17:05:01.432214 10926 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00297231 (* 0.0909091 = 0.00027021 loss) | |
| I0510 17:05:01.432227 10926 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00232996 (* 0.0909091 = 0.000211815 loss) | |
| I0510 17:05:01.432241 10926 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00107881 (* 0.0909091 = 9.80735e-05 loss) | |
| I0510 17:05:01.432255 10926 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00098225 (* 0.0909091 = 8.92954e-05 loss) | |
| I0510 17:05:01.432271 10926 solver.cpp:245] Train net output #146: loss3/loss22 = 0.000996036 (* 0.0909091 = 9.05487e-05 loss) | |
| I0510 17:05:01.432282 10926 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0510 17:05:01.432298 10926 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0510 17:05:01.432310 10926 solver.cpp:245] Train net output #149: total_confidence = 0.000274699 | |
| I0510 17:05:01.432332 10926 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.000377441 | |
| I0510 17:05:01.432348 10926 sgd_solver.cpp:106] Iteration 19500, lr = 0.001 | |
| I0510 17:06:27.957352 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 31.6958 > 30) by scale factor 0.946497 | |
| I0510 17:07:29.177783 10926 solver.cpp:456] Snapshotting to binary proto file /mnt/snapshots/mixed_lstm15_bn_iter_20000.caffemodel | |
| I0510 17:07:29.625097 10926 sgd_solver.cpp:273] Snapshotting solver state to binary proto file /mnt/snapshots/mixed_lstm15_bn_iter_20000.solverstate | |
| I0510 17:07:29.845387 10926 solver.cpp:338] Iteration 20000, Testing net (#0) | |
| I0510 17:08:13.388980 10926 solver.cpp:393] Test loss: 9.12062 | |
| I0510 17:08:13.389135 10926 solver.cpp:406] Test net output #0: loss1/accuracy = 0.065402 | |
| I0510 17:08:13.389158 10926 solver.cpp:406] Test net output #1: loss1/accuracy01 = 0.124 | |
| I0510 17:08:13.389173 10926 solver.cpp:406] Test net output #2: loss1/accuracy02 = 0.101 | |
| I0510 17:08:13.389186 10926 solver.cpp:406] Test net output #3: loss1/accuracy03 = 0.074 | |
| I0510 17:08:13.389199 10926 solver.cpp:406] Test net output #4: loss1/accuracy04 = 0.167 | |
| I0510 17:08:13.389211 10926 solver.cpp:406] Test net output #5: loss1/accuracy05 = 0.329 | |
| I0510 17:08:13.389224 10926 solver.cpp:406] Test net output #6: loss1/accuracy06 = 0.46 | |
| I0510 17:08:13.389236 10926 solver.cpp:406] Test net output #7: loss1/accuracy07 = 0.718 | |
| I0510 17:08:13.389248 10926 solver.cpp:406] Test net output #8: loss1/accuracy08 = 0.906 | |
| I0510 17:08:13.389261 10926 solver.cpp:406] Test net output #9: loss1/accuracy09 = 0.988 | |
| I0510 17:08:13.389273 10926 solver.cpp:406] Test net output #10: loss1/accuracy10 = 0.998 | |
| I0510 17:08:13.389286 10926 solver.cpp:406] Test net output #11: loss1/accuracy11 = 1 | |
| I0510 17:08:13.389298 10926 solver.cpp:406] Test net output #12: loss1/accuracy12 = 1 | |
| I0510 17:08:13.389310 10926 solver.cpp:406] Test net output #13: loss1/accuracy13 = 1 | |
| I0510 17:08:13.389322 10926 solver.cpp:406] Test net output #14: loss1/accuracy14 = 1 | |
| I0510 17:08:13.389333 10926 solver.cpp:406] Test net output #15: loss1/accuracy15 = 1 | |
| I0510 17:08:13.389345 10926 solver.cpp:406] Test net output #16: loss1/accuracy16 = 1 | |
| I0510 17:08:13.389356 10926 solver.cpp:406] Test net output #17: loss1/accuracy17 = 1 | |
| I0510 17:08:13.389367 10926 solver.cpp:406] Test net output #18: loss1/accuracy18 = 1 | |
| I0510 17:08:13.389379 10926 solver.cpp:406] Test net output #19: loss1/accuracy19 = 1 | |
| I0510 17:08:13.389391 10926 solver.cpp:406] Test net output #20: loss1/accuracy20 = 1 | |
| I0510 17:08:13.389402 10926 solver.cpp:406] Test net output #21: loss1/accuracy21 = 1 | |
| I0510 17:08:13.389415 10926 solver.cpp:406] Test net output #22: loss1/accuracy22 = 1 | |
| I0510 17:08:13.389426 10926 solver.cpp:406] Test net output #23: loss1/accuracy_incl_empty = 0.763683 | |
| I0510 17:08:13.389438 10926 solver.cpp:406] Test net output #24: loss1/accuracy_top3 = 0.235184 | |
| I0510 17:08:13.389456 10926 solver.cpp:406] Test net output #25: loss1/cross_entropy_loss = 3.46332 (* 0.3 = 1.039 loss) | |
| I0510 17:08:13.389470 10926 solver.cpp:406] Test net output #26: loss1/cross_entropy_loss_incl_empty = 0.904009 (* 0.3 = 0.271203 loss) | |
| I0510 17:08:13.389485 10926 solver.cpp:406] Test net output #27: loss1/loss01 = 2.94596 (* 0.0272727 = 0.0803444 loss) | |
| I0510 17:08:13.389499 10926 solver.cpp:406] Test net output #28: loss1/loss02 = 3.11771 (* 0.0272727 = 0.0850284 loss) | |
| I0510 17:08:13.389513 10926 solver.cpp:406] Test net output #29: loss1/loss03 = 3.19968 (* 0.0272727 = 0.087264 loss) | |
| I0510 17:08:13.389528 10926 solver.cpp:406] Test net output #30: loss1/loss04 = 3.0113 (* 0.0272727 = 0.0821264 loss) | |
| I0510 17:08:13.389542 10926 solver.cpp:406] Test net output #31: loss1/loss05 = 2.55598 (* 0.0272727 = 0.0697084 loss) | |
| I0510 17:08:13.389556 10926 solver.cpp:406] Test net output #32: loss1/loss06 = 2.15753 (* 0.0272727 = 0.0588418 loss) | |
| I0510 17:08:13.389576 10926 solver.cpp:406] Test net output #33: loss1/loss07 = 1.36372 (* 0.0272727 = 0.0371924 loss) | |
| I0510 17:08:13.389590 10926 solver.cpp:406] Test net output #34: loss1/loss08 = 0.507759 (* 0.0272727 = 0.013848 loss) | |
| I0510 17:08:13.389605 10926 solver.cpp:406] Test net output #35: loss1/loss09 = 0.106403 (* 0.0272727 = 0.0029019 loss) | |
| I0510 17:08:13.389619 10926 solver.cpp:406] Test net output #36: loss1/loss10 = 0.0434494 (* 0.0272727 = 0.00118498 loss) | |
| I0510 17:08:13.389634 10926 solver.cpp:406] Test net output #37: loss1/loss11 = 0.0294678 (* 0.0272727 = 0.000803668 loss) | |
| I0510 17:08:13.389657 10926 solver.cpp:406] Test net output #38: loss1/loss12 = 0.022369 (* 0.0272727 = 0.000610064 loss) | |
| I0510 17:08:13.389672 10926 solver.cpp:406] Test net output #39: loss1/loss13 = 0.0174541 (* 0.0272727 = 0.000476022 loss) | |
| I0510 17:08:13.389708 10926 solver.cpp:406] Test net output #40: loss1/loss14 = 0.0139357 (* 0.0272727 = 0.000380064 loss) | |
| I0510 17:08:13.389724 10926 solver.cpp:406] Test net output #41: loss1/loss15 = 0.0105754 (* 0.0272727 = 0.000288419 loss) | |
| I0510 17:08:13.389737 10926 solver.cpp:406] Test net output #42: loss1/loss16 = 0.00818925 (* 0.0272727 = 0.000223343 loss) | |
| I0510 17:08:13.389751 10926 solver.cpp:406] Test net output #43: loss1/loss17 = 0.00434872 (* 0.0272727 = 0.000118601 loss) | |
| I0510 17:08:13.389765 10926 solver.cpp:406] Test net output #44: loss1/loss18 = 0.00341605 (* 0.0272727 = 9.3165e-05 loss) | |
| I0510 17:08:13.389780 10926 solver.cpp:406] Test net output #45: loss1/loss19 = 0.00315858 (* 0.0272727 = 8.6143e-05 loss) | |
| I0510 17:08:13.389793 10926 solver.cpp:406] Test net output #46: loss1/loss20 = 0.00277439 (* 0.0272727 = 7.56651e-05 loss) | |
| I0510 17:08:13.389807 10926 solver.cpp:406] Test net output #47: loss1/loss21 = 0.00259239 (* 0.0272727 = 7.07015e-05 loss) | |
| I0510 17:08:13.389822 10926 solver.cpp:406] Test net output #48: loss1/loss22 = 0.00251618 (* 0.0272727 = 6.86232e-05 loss) | |
| I0510 17:08:13.389833 10926 solver.cpp:406] Test net output #49: loss2/accuracy = 0.0674667 | |
| I0510 17:08:13.389847 10926 solver.cpp:406] Test net output #50: loss2/accuracy01 = 0.124 | |
| I0510 17:08:13.389858 10926 solver.cpp:406] Test net output #51: loss2/accuracy02 = 0.086 | |
| I0510 17:08:13.389870 10926 solver.cpp:406] Test net output #52: loss2/accuracy03 = 0.088 | |
| I0510 17:08:13.389886 10926 solver.cpp:406] Test net output #53: loss2/accuracy04 = 0.174 | |
| I0510 17:08:13.389899 10926 solver.cpp:406] Test net output #54: loss2/accuracy05 = 0.329 | |
| I0510 17:08:13.389909 10926 solver.cpp:406] Test net output #55: loss2/accuracy06 = 0.457 | |
| I0510 17:08:13.389922 10926 solver.cpp:406] Test net output #56: loss2/accuracy07 = 0.717 | |
| I0510 17:08:13.389933 10926 solver.cpp:406] Test net output #57: loss2/accuracy08 = 0.905 | |
| I0510 17:08:13.389945 10926 solver.cpp:406] Test net output #58: loss2/accuracy09 = 0.988 | |
| I0510 17:08:13.389956 10926 solver.cpp:406] Test net output #59: loss2/accuracy10 = 0.999 | |
| I0510 17:08:13.389969 10926 solver.cpp:406] Test net output #60: loss2/accuracy11 = 1 | |
| I0510 17:08:13.389981 10926 solver.cpp:406] Test net output #61: loss2/accuracy12 = 1 | |
| I0510 17:08:13.389992 10926 solver.cpp:406] Test net output #62: loss2/accuracy13 = 1 | |
| I0510 17:08:13.390003 10926 solver.cpp:406] Test net output #63: loss2/accuracy14 = 1 | |
| I0510 17:08:13.390014 10926 solver.cpp:406] Test net output #64: loss2/accuracy15 = 1 | |
| I0510 17:08:13.390027 10926 solver.cpp:406] Test net output #65: loss2/accuracy16 = 1 | |
| I0510 17:08:13.390038 10926 solver.cpp:406] Test net output #66: loss2/accuracy17 = 1 | |
| I0510 17:08:13.390048 10926 solver.cpp:406] Test net output #67: loss2/accuracy18 = 1 | |
| I0510 17:08:13.390060 10926 solver.cpp:406] Test net output #68: loss2/accuracy19 = 1 | |
| I0510 17:08:13.390071 10926 solver.cpp:406] Test net output #69: loss2/accuracy20 = 1 | |
| I0510 17:08:13.390082 10926 solver.cpp:406] Test net output #70: loss2/accuracy21 = 1 | |
| I0510 17:08:13.390094 10926 solver.cpp:406] Test net output #71: loss2/accuracy22 = 1 | |
| I0510 17:08:13.390105 10926 solver.cpp:406] Test net output #72: loss2/accuracy_incl_empty = 0.763774 | |
| I0510 17:08:13.390117 10926 solver.cpp:406] Test net output #73: loss2/accuracy_top3 = 0.23443 | |
| I0510 17:08:13.390130 10926 solver.cpp:406] Test net output #74: loss2/cross_entropy_loss = 3.44247 (* 0.3 = 1.03274 loss) | |
| I0510 17:08:13.390147 10926 solver.cpp:406] Test net output #75: loss2/cross_entropy_loss_incl_empty = 0.904283 (* 0.3 = 0.271285 loss) | |
| I0510 17:08:13.390162 10926 solver.cpp:406] Test net output #76: loss2/loss01 = 2.92581 (* 0.0272727 = 0.0797948 loss) | |
| I0510 17:08:13.390177 10926 solver.cpp:406] Test net output #77: loss2/loss02 = 3.07592 (* 0.0272727 = 0.0838888 loss) | |
| I0510 17:08:13.390189 10926 solver.cpp:406] Test net output #78: loss2/loss03 = 3.16184 (* 0.0272727 = 0.086232 loss) | |
| I0510 17:08:13.390215 10926 solver.cpp:406] Test net output #79: loss2/loss04 = 2.97811 (* 0.0272727 = 0.0812212 loss) | |
| I0510 17:08:13.390230 10926 solver.cpp:406] Test net output #80: loss2/loss05 = 2.53885 (* 0.0272727 = 0.0692413 loss) | |
| I0510 17:08:13.390241 10926 solver.cpp:406] Test net output #81: loss2/loss06 = 2.141 (* 0.0272727 = 0.058391 loss) | |
| I0510 17:08:13.390250 10926 solver.cpp:406] Test net output #82: loss2/loss07 = 1.33616 (* 0.0272727 = 0.0364407 loss) | |
| I0510 17:08:13.390264 10926 solver.cpp:406] Test net output #83: loss2/loss08 = 0.503014 (* 0.0272727 = 0.0137186 loss) | |
| I0510 17:08:13.390278 10926 solver.cpp:406] Test net output #84: loss2/loss09 = 0.106417 (* 0.0272727 = 0.00290229 loss) | |
| I0510 17:08:13.390293 10926 solver.cpp:406] Test net output #85: loss2/loss10 = 0.0399665 (* 0.0272727 = 0.00108999 loss) | |
| I0510 17:08:13.390307 10926 solver.cpp:406] Test net output #86: loss2/loss11 = 0.0266049 (* 0.0272727 = 0.000725589 loss) | |
| I0510 17:08:13.390321 10926 solver.cpp:406] Test net output #87: loss2/loss12 = 0.0180607 (* 0.0272727 = 0.000492563 loss) | |
| I0510 17:08:13.390339 10926 solver.cpp:406] Test net output #88: loss2/loss13 = 0.0138763 (* 0.0272727 = 0.000378444 loss) | |
| I0510 17:08:13.390353 10926 solver.cpp:406] Test net output #89: loss2/loss14 = 0.0110355 (* 0.0272727 = 0.000300969 loss) | |
| I0510 17:08:13.390367 10926 solver.cpp:406] Test net output #90: loss2/loss15 = 0.00830702 (* 0.0272727 = 0.000226555 loss) | |
| I0510 17:08:13.390382 10926 solver.cpp:406] Test net output #91: loss2/loss16 = 0.00607517 (* 0.0272727 = 0.000165686 loss) | |
| I0510 17:08:13.390395 10926 solver.cpp:406] Test net output #92: loss2/loss17 = 0.0033327 (* 0.0272727 = 9.08919e-05 loss) | |
| I0510 17:08:13.390409 10926 solver.cpp:406] Test net output #93: loss2/loss18 = 0.00311261 (* 0.0272727 = 8.48894e-05 loss) | |
| I0510 17:08:13.390424 10926 solver.cpp:406] Test net output #94: loss2/loss19 = 0.00227449 (* 0.0272727 = 6.20317e-05 loss) | |
| I0510 17:08:13.390436 10926 solver.cpp:406] Test net output #95: loss2/loss20 = 0.00239422 (* 0.0272727 = 6.52968e-05 loss) | |
| I0510 17:08:13.390450 10926 solver.cpp:406] Test net output #96: loss2/loss21 = 0.00210383 (* 0.0272727 = 5.73771e-05 loss) | |
| I0510 17:08:13.390463 10926 solver.cpp:406] Test net output #97: loss2/loss22 = 0.00234059 (* 0.0272727 = 6.38343e-05 loss) | |
| I0510 17:08:13.390483 10926 solver.cpp:406] Test net output #98: loss3/accuracy = 0.0969056 | |
| I0510 17:08:13.390496 10926 solver.cpp:406] Test net output #99: loss3/accuracy01 = 0.119 | |
| I0510 17:08:13.390507 10926 solver.cpp:406] Test net output #100: loss3/accuracy02 = 0.095 | |
| I0510 17:08:13.390519 10926 solver.cpp:406] Test net output #101: loss3/accuracy03 = 0.095 | |
| I0510 17:08:13.390530 10926 solver.cpp:406] Test net output #102: loss3/accuracy04 = 0.167 | |
| I0510 17:08:13.390544 10926 solver.cpp:406] Test net output #103: loss3/accuracy05 = 0.326 | |
| I0510 17:08:13.390557 10926 solver.cpp:406] Test net output #104: loss3/accuracy06 = 0.452 | |
| I0510 17:08:13.390568 10926 solver.cpp:406] Test net output #105: loss3/accuracy07 = 0.714 | |
| I0510 17:08:13.390579 10926 solver.cpp:406] Test net output #106: loss3/accuracy08 = 0.902 | |
| I0510 17:08:13.390590 10926 solver.cpp:406] Test net output #107: loss3/accuracy09 = 0.985 | |
| I0510 17:08:13.390607 10926 solver.cpp:406] Test net output #108: loss3/accuracy10 = 0.998 | |
| I0510 17:08:13.390619 10926 solver.cpp:406] Test net output #109: loss3/accuracy11 = 0.999 | |
| I0510 17:08:13.390630 10926 solver.cpp:406] Test net output #110: loss3/accuracy12 = 0.999 | |
| I0510 17:08:13.390642 10926 solver.cpp:406] Test net output #111: loss3/accuracy13 = 0.999 | |
| I0510 17:08:13.390653 10926 solver.cpp:406] Test net output #112: loss3/accuracy14 = 1 | |
| I0510 17:08:13.390672 10926 solver.cpp:406] Test net output #113: loss3/accuracy15 = 1 | |
| I0510 17:08:13.390683 10926 solver.cpp:406] Test net output #114: loss3/accuracy16 = 1 | |
| I0510 17:08:13.390704 10926 solver.cpp:406] Test net output #115: loss3/accuracy17 = 1 | |
| I0510 17:08:13.390717 10926 solver.cpp:406] Test net output #116: loss3/accuracy18 = 1 | |
| I0510 17:08:13.390728 10926 solver.cpp:406] Test net output #117: loss3/accuracy19 = 1 | |
| I0510 17:08:13.390739 10926 solver.cpp:406] Test net output #118: loss3/accuracy20 = 1 | |
| I0510 17:08:13.390750 10926 solver.cpp:406] Test net output #119: loss3/accuracy21 = 1 | |
| I0510 17:08:13.390763 10926 solver.cpp:406] Test net output #120: loss3/accuracy22 = 1 | |
| I0510 17:08:13.390774 10926 solver.cpp:406] Test net output #121: loss3/accuracy_incl_empty = 0.766228 | |
| I0510 17:08:13.390785 10926 solver.cpp:406] Test net output #122: loss3/accuracy_top3 = 0.268275 | |
| I0510 17:08:13.390799 10926 solver.cpp:406] Test net output #123: loss3/cross_entropy_loss = 3.00705 (* 1 = 3.00705 loss) | |
| I0510 17:08:13.390812 10926 solver.cpp:406] Test net output #124: loss3/cross_entropy_loss_incl_empty = 0.828076 (* 1 = 0.828076 loss) | |
| I0510 17:08:13.390826 10926 solver.cpp:406] Test net output #125: loss3/loss01 = 2.7835 (* 0.0909091 = 0.253045 loss) | |
| I0510 17:08:13.390841 10926 solver.cpp:406] Test net output #126: loss3/loss02 = 2.97154 (* 0.0909091 = 0.27014 loss) | |
| I0510 17:08:13.390853 10926 solver.cpp:406] Test net output #127: loss3/loss03 = 3.02609 (* 0.0909091 = 0.275099 loss) | |
| I0510 17:08:13.390867 10926 solver.cpp:406] Test net output #128: loss3/loss04 = 2.87465 (* 0.0909091 = 0.261332 loss) | |
| I0510 17:08:13.390882 10926 solver.cpp:406] Test net output #129: loss3/loss05 = 2.40703 (* 0.0909091 = 0.218821 loss) | |
| I0510 17:08:13.390895 10926 solver.cpp:406] Test net output #130: loss3/loss06 = 2.0175 (* 0.0909091 = 0.183409 loss) | |
| I0510 17:08:13.390909 10926 solver.cpp:406] Test net output #131: loss3/loss07 = 1.22537 (* 0.0909091 = 0.111397 loss) | |
| I0510 17:08:13.390923 10926 solver.cpp:406] Test net output #132: loss3/loss08 = 0.454826 (* 0.0909091 = 0.0413478 loss) | |
| I0510 17:08:13.390939 10926 solver.cpp:406] Test net output #133: loss3/loss09 = 0.103141 (* 0.0909091 = 0.00937641 loss) | |
| I0510 17:08:13.390954 10926 solver.cpp:406] Test net output #134: loss3/loss10 = 0.0367448 (* 0.0909091 = 0.00334044 loss) | |
| I0510 17:08:13.390967 10926 solver.cpp:406] Test net output #135: loss3/loss11 = 0.0229749 (* 0.0909091 = 0.00208863 loss) | |
| I0510 17:08:13.390981 10926 solver.cpp:406] Test net output #136: loss3/loss12 = 0.0150689 (* 0.0909091 = 0.0013699 loss) | |
| I0510 17:08:13.390996 10926 solver.cpp:406] Test net output #137: loss3/loss13 = 0.0106876 (* 0.0909091 = 0.000971603 loss) | |
| I0510 17:08:13.391010 10926 solver.cpp:406] Test net output #138: loss3/loss14 = 0.00795318 (* 0.0909091 = 0.000723016 loss) | |
| I0510 17:08:13.391023 10926 solver.cpp:406] Test net output #139: loss3/loss15 = 0.00556155 (* 0.0909091 = 0.000505595 loss) | |
| I0510 17:08:13.391037 10926 solver.cpp:406] Test net output #140: loss3/loss16 = 0.00417096 (* 0.0909091 = 0.000379178 loss) | |
| I0510 17:08:13.391052 10926 solver.cpp:406] Test net output #141: loss3/loss17 = 0.0018239 (* 0.0909091 = 0.000165809 loss) | |
| I0510 17:08:13.391065 10926 solver.cpp:406] Test net output #142: loss3/loss18 = 0.00129458 (* 0.0909091 = 0.000117689 loss) | |
| I0510 17:08:13.391079 10926 solver.cpp:406] Test net output #143: loss3/loss19 = 0.00091107 (* 0.0909091 = 8.28245e-05 loss) | |
| I0510 17:08:13.391093 10926 solver.cpp:406] Test net output #144: loss3/loss20 = 0.000791456 (* 0.0909091 = 7.19506e-05 loss) | |
| I0510 17:08:13.391106 10926 solver.cpp:406] Test net output #145: loss3/loss21 = 0.000705039 (* 0.0909091 = 6.40945e-05 loss) | |
| I0510 17:08:13.391120 10926 solver.cpp:406] Test net output #146: loss3/loss22 = 0.000568027 (* 0.0909091 = 5.16389e-05 loss) | |
| I0510 17:08:13.391132 10926 solver.cpp:406] Test net output #147: total_accuracy = 0 | |
| I0510 17:08:13.391144 10926 solver.cpp:406] Test net output #148: total_accuracy_not_rec = 0 | |
| I0510 17:08:13.391155 10926 solver.cpp:406] Test net output #149: total_confidence = 0.000113181 | |
| I0510 17:08:13.391166 10926 solver.cpp:406] Test net output #150: total_confidence_not_rec = 0.000148959 | |
| I0510 17:08:13.391191 10926 solver.cpp:338] Iteration 20000, Testing net (#1) | |
| I0510 17:08:56.572163 10926 solver.cpp:393] Test loss: 9.77909 | |
| I0510 17:08:56.572270 10926 solver.cpp:406] Test net output #0: loss1/accuracy = 0.0662082 | |
| I0510 17:08:56.572290 10926 solver.cpp:406] Test net output #1: loss1/accuracy01 = 0.134 | |
| I0510 17:08:56.572305 10926 solver.cpp:406] Test net output #2: loss1/accuracy02 = 0.102 | |
| I0510 17:08:56.572319 10926 solver.cpp:406] Test net output #3: loss1/accuracy03 = 0.084 | |
| I0510 17:08:56.572331 10926 solver.cpp:406] Test net output #4: loss1/accuracy04 = 0.158 | |
| I0510 17:08:56.572345 10926 solver.cpp:406] Test net output #5: loss1/accuracy05 = 0.319 | |
| I0510 17:08:56.572356 10926 solver.cpp:406] Test net output #6: loss1/accuracy06 = 0.435 | |
| I0510 17:08:56.572370 10926 solver.cpp:406] Test net output #7: loss1/accuracy07 = 0.67 | |
| I0510 17:08:56.572381 10926 solver.cpp:406] Test net output #8: loss1/accuracy08 = 0.82 | |
| I0510 17:08:56.572394 10926 solver.cpp:406] Test net output #9: loss1/accuracy09 = 0.906 | |
| I0510 17:08:56.572413 10926 solver.cpp:406] Test net output #10: loss1/accuracy10 = 0.918 | |
| I0510 17:08:56.572427 10926 solver.cpp:406] Test net output #11: loss1/accuracy11 = 0.937 | |
| I0510 17:08:56.572438 10926 solver.cpp:406] Test net output #12: loss1/accuracy12 = 0.946 | |
| I0510 17:08:56.572451 10926 solver.cpp:406] Test net output #13: loss1/accuracy13 = 0.96 | |
| I0510 17:08:56.572463 10926 solver.cpp:406] Test net output #14: loss1/accuracy14 = 0.971 | |
| I0510 17:08:56.572484 10926 solver.cpp:406] Test net output #15: loss1/accuracy15 = 0.978 | |
| I0510 17:08:56.572495 10926 solver.cpp:406] Test net output #16: loss1/accuracy16 = 0.984 | |
| I0510 17:08:56.572509 10926 solver.cpp:406] Test net output #17: loss1/accuracy17 = 0.994 | |
| I0510 17:08:56.572520 10926 solver.cpp:406] Test net output #18: loss1/accuracy18 = 0.996 | |
| I0510 17:08:56.572532 10926 solver.cpp:406] Test net output #19: loss1/accuracy19 = 0.997 | |
| I0510 17:08:56.572545 10926 solver.cpp:406] Test net output #20: loss1/accuracy20 = 0.997 | |
| I0510 17:08:56.572556 10926 solver.cpp:406] Test net output #21: loss1/accuracy21 = 0.998 | |
| I0510 17:08:56.572568 10926 solver.cpp:406] Test net output #22: loss1/accuracy22 = 0.999 | |
| I0510 17:08:56.572581 10926 solver.cpp:406] Test net output #23: loss1/accuracy_incl_empty = 0.735591 | |
| I0510 17:08:56.572592 10926 solver.cpp:406] Test net output #24: loss1/accuracy_top3 = 0.238636 | |
| I0510 17:08:56.572609 10926 solver.cpp:406] Test net output #25: loss1/cross_entropy_loss = 3.48972 (* 0.3 = 1.04692 loss) | |
| I0510 17:08:56.572633 10926 solver.cpp:406] Test net output #26: loss1/cross_entropy_loss_incl_empty = 1.02404 (* 0.3 = 0.307211 loss) | |
| I0510 17:08:56.572648 10926 solver.cpp:406] Test net output #27: loss1/loss01 = 2.95389 (* 0.0272727 = 0.0805608 loss) | |
| I0510 17:08:56.572661 10926 solver.cpp:406] Test net output #28: loss1/loss02 = 3.11736 (* 0.0272727 = 0.085019 loss) | |
| I0510 17:08:56.572674 10926 solver.cpp:406] Test net output #29: loss1/loss03 = 3.22915 (* 0.0272727 = 0.0880678 loss) | |
| I0510 17:08:56.572688 10926 solver.cpp:406] Test net output #30: loss1/loss04 = 3.12008 (* 0.0272727 = 0.085093 loss) | |
| I0510 17:08:56.572710 10926 solver.cpp:406] Test net output #31: loss1/loss05 = 2.69261 (* 0.0272727 = 0.0734348 loss) | |
| I0510 17:08:56.572723 10926 solver.cpp:406] Test net output #32: loss1/loss06 = 2.29093 (* 0.0272727 = 0.0624798 loss) | |
| I0510 17:08:56.572737 10926 solver.cpp:406] Test net output #33: loss1/loss07 = 1.50802 (* 0.0272727 = 0.0411278 loss) | |
| I0510 17:08:56.572751 10926 solver.cpp:406] Test net output #34: loss1/loss08 = 0.847533 (* 0.0272727 = 0.0231145 loss) | |
| I0510 17:08:56.572764 10926 solver.cpp:406] Test net output #35: loss1/loss09 = 0.439358 (* 0.0272727 = 0.0119825 loss) | |
| I0510 17:08:56.572778 10926 solver.cpp:406] Test net output #36: loss1/loss10 = 0.370485 (* 0.0272727 = 0.0101041 loss) | |
| I0510 17:08:56.572793 10926 solver.cpp:406] Test net output #37: loss1/loss11 = 0.300796 (* 0.0272727 = 0.00820352 loss) | |
| I0510 17:08:56.572806 10926 solver.cpp:406] Test net output #38: loss1/loss12 = 0.276369 (* 0.0272727 = 0.00753733 loss) | |
| I0510 17:08:56.572841 10926 solver.cpp:406] Test net output #39: loss1/loss13 = 0.213319 (* 0.0272727 = 0.0058178 loss) | |
| I0510 17:08:56.572857 10926 solver.cpp:406] Test net output #40: loss1/loss14 = 0.164338 (* 0.0272727 = 0.00448195 loss) | |
| I0510 17:08:56.572871 10926 solver.cpp:406] Test net output #41: loss1/loss15 = 0.129904 (* 0.0272727 = 0.00354283 loss) | |
| I0510 17:08:56.572890 10926 solver.cpp:406] Test net output #42: loss1/loss16 = 0.104743 (* 0.0272727 = 0.00285664 loss) | |
| I0510 17:08:56.572903 10926 solver.cpp:406] Test net output #43: loss1/loss17 = 0.0462588 (* 0.0272727 = 0.0012616 loss) | |
| I0510 17:08:56.572918 10926 solver.cpp:406] Test net output #44: loss1/loss18 = 0.0356722 (* 0.0272727 = 0.000972878 loss) | |
| I0510 17:08:56.572932 10926 solver.cpp:406] Test net output #45: loss1/loss19 = 0.0271005 (* 0.0272727 = 0.000739103 loss) | |
| I0510 17:08:56.572947 10926 solver.cpp:406] Test net output #46: loss1/loss20 = 0.0272262 (* 0.0272727 = 0.000742533 loss) | |
| I0510 17:08:56.572960 10926 solver.cpp:406] Test net output #47: loss1/loss21 = 0.0194593 (* 0.0272727 = 0.000530708 loss) | |
| I0510 17:08:56.572973 10926 solver.cpp:406] Test net output #48: loss1/loss22 = 0.0120397 (* 0.0272727 = 0.000328355 loss) | |
| I0510 17:08:56.572986 10926 solver.cpp:406] Test net output #49: loss2/accuracy = 0.0709399 | |
| I0510 17:08:56.572999 10926 solver.cpp:406] Test net output #50: loss2/accuracy01 = 0.132 | |
| I0510 17:08:56.573010 10926 solver.cpp:406] Test net output #51: loss2/accuracy02 = 0.095 | |
| I0510 17:08:56.573022 10926 solver.cpp:406] Test net output #52: loss2/accuracy03 = 0.081 | |
| I0510 17:08:56.573036 10926 solver.cpp:406] Test net output #53: loss2/accuracy04 = 0.166 | |
| I0510 17:08:56.573048 10926 solver.cpp:406] Test net output #54: loss2/accuracy05 = 0.319 | |
| I0510 17:08:56.573060 10926 solver.cpp:406] Test net output #55: loss2/accuracy06 = 0.433 | |
| I0510 17:08:56.573072 10926 solver.cpp:406] Test net output #56: loss2/accuracy07 = 0.673 | |
| I0510 17:08:56.573083 10926 solver.cpp:406] Test net output #57: loss2/accuracy08 = 0.818 | |
| I0510 17:08:56.573096 10926 solver.cpp:406] Test net output #58: loss2/accuracy09 = 0.905 | |
| I0510 17:08:56.573107 10926 solver.cpp:406] Test net output #59: loss2/accuracy10 = 0.918 | |
| I0510 17:08:56.573137 10926 solver.cpp:406] Test net output #60: loss2/accuracy11 = 0.938 | |
| I0510 17:08:56.573150 10926 solver.cpp:406] Test net output #61: loss2/accuracy12 = 0.946 | |
| I0510 17:08:56.573159 10926 solver.cpp:406] Test net output #62: loss2/accuracy13 = 0.96 | |
| I0510 17:08:56.573168 10926 solver.cpp:406] Test net output #63: loss2/accuracy14 = 0.971 | |
| I0510 17:08:56.573179 10926 solver.cpp:406] Test net output #64: loss2/accuracy15 = 0.978 | |
| I0510 17:08:56.573192 10926 solver.cpp:406] Test net output #65: loss2/accuracy16 = 0.984 | |
| I0510 17:08:56.573204 10926 solver.cpp:406] Test net output #66: loss2/accuracy17 = 0.994 | |
| I0510 17:08:56.573216 10926 solver.cpp:406] Test net output #67: loss2/accuracy18 = 0.996 | |
| I0510 17:08:56.573228 10926 solver.cpp:406] Test net output #68: loss2/accuracy19 = 0.997 | |
| I0510 17:08:56.573241 10926 solver.cpp:406] Test net output #69: loss2/accuracy20 = 0.997 | |
| I0510 17:08:56.573252 10926 solver.cpp:406] Test net output #70: loss2/accuracy21 = 0.998 | |
| I0510 17:08:56.573264 10926 solver.cpp:406] Test net output #71: loss2/accuracy22 = 0.999 | |
| I0510 17:08:56.573276 10926 solver.cpp:406] Test net output #72: loss2/accuracy_incl_empty = 0.736274 | |
| I0510 17:08:56.573287 10926 solver.cpp:406] Test net output #73: loss2/accuracy_top3 = 0.239177 | |
| I0510 17:08:56.573305 10926 solver.cpp:406] Test net output #74: loss2/cross_entropy_loss = 3.49565 (* 0.3 = 1.0487 loss) | |
| I0510 17:08:56.573319 10926 solver.cpp:406] Test net output #75: loss2/cross_entropy_loss_incl_empty = 1.03395 (* 0.3 = 0.310186 loss) | |
| I0510 17:08:56.573333 10926 solver.cpp:406] Test net output #76: loss2/loss01 = 2.92867 (* 0.0272727 = 0.0798728 loss) | |
| I0510 17:08:56.573348 10926 solver.cpp:406] Test net output #77: loss2/loss02 = 3.08853 (* 0.0272727 = 0.0842326 loss) | |
| I0510 17:08:56.573374 10926 solver.cpp:406] Test net output #78: loss2/loss03 = 3.21149 (* 0.0272727 = 0.087586 loss) | |
| I0510 17:08:56.573390 10926 solver.cpp:406] Test net output #79: loss2/loss04 = 3.08678 (* 0.0272727 = 0.084185 loss) | |
| I0510 17:08:56.573408 10926 solver.cpp:406] Test net output #80: loss2/loss05 = 2.67171 (* 0.0272727 = 0.0728648 loss) | |
| I0510 17:08:56.573422 10926 solver.cpp:406] Test net output #81: loss2/loss06 = 2.27801 (* 0.0272727 = 0.0621275 loss) | |
| I0510 17:08:56.573436 10926 solver.cpp:406] Test net output #82: loss2/loss07 = 1.49141 (* 0.0272727 = 0.0406748 loss) | |
| I0510 17:08:56.573451 10926 solver.cpp:406] Test net output #83: loss2/loss08 = 0.83178 (* 0.0272727 = 0.0226849 loss) | |
| I0510 17:08:56.573463 10926 solver.cpp:406] Test net output #84: loss2/loss09 = 0.434707 (* 0.0272727 = 0.0118556 loss) | |
| I0510 17:08:56.573483 10926 solver.cpp:406] Test net output #85: loss2/loss10 = 0.363094 (* 0.0272727 = 0.00990256 loss) | |
| I0510 17:08:56.573498 10926 solver.cpp:406] Test net output #86: loss2/loss11 = 0.296223 (* 0.0272727 = 0.00807882 loss) | |
| I0510 17:08:56.573511 10926 solver.cpp:406] Test net output #87: loss2/loss12 = 0.273676 (* 0.0272727 = 0.00746388 loss) | |
| I0510 17:08:56.573525 10926 solver.cpp:406] Test net output #88: loss2/loss13 = 0.210088 (* 0.0272727 = 0.00572967 loss) | |
| I0510 17:08:56.573539 10926 solver.cpp:406] Test net output #89: loss2/loss14 = 0.163255 (* 0.0272727 = 0.00445241 loss) | |
| I0510 17:08:56.573552 10926 solver.cpp:406] Test net output #90: loss2/loss15 = 0.1285 (* 0.0272727 = 0.00350454 loss) | |
| I0510 17:08:56.573566 10926 solver.cpp:406] Test net output #91: loss2/loss16 = 0.105504 (* 0.0272727 = 0.00287739 loss) | |
| I0510 17:08:56.573580 10926 solver.cpp:406] Test net output #92: loss2/loss17 = 0.0483667 (* 0.0272727 = 0.00131909 loss) | |
| I0510 17:08:56.573593 10926 solver.cpp:406] Test net output #93: loss2/loss18 = 0.0369725 (* 0.0272727 = 0.00100834 loss) | |
| I0510 17:08:56.573607 10926 solver.cpp:406] Test net output #94: loss2/loss19 = 0.0285367 (* 0.0272727 = 0.000778275 loss) | |
| I0510 17:08:56.573621 10926 solver.cpp:406] Test net output #95: loss2/loss20 = 0.0293199 (* 0.0272727 = 0.000799634 loss) | |
| I0510 17:08:56.573634 10926 solver.cpp:406] Test net output #96: loss2/loss21 = 0.0221133 (* 0.0272727 = 0.000603089 loss) | |
| I0510 17:08:56.573648 10926 solver.cpp:406] Test net output #97: loss2/loss22 = 0.01207 (* 0.0272727 = 0.000329181 loss) | |
| I0510 17:08:56.573659 10926 solver.cpp:406] Test net output #98: loss3/accuracy = 0.0984889 | |
| I0510 17:08:56.573671 10926 solver.cpp:406] Test net output #99: loss3/accuracy01 = 0.122 | |
| I0510 17:08:56.573683 10926 solver.cpp:406] Test net output #100: loss3/accuracy02 = 0.097 | |
| I0510 17:08:56.573695 10926 solver.cpp:406] Test net output #101: loss3/accuracy03 = 0.097 | |
| I0510 17:08:56.573706 10926 solver.cpp:406] Test net output #102: loss3/accuracy04 = 0.153 | |
| I0510 17:08:56.573719 10926 solver.cpp:406] Test net output #103: loss3/accuracy05 = 0.319 | |
| I0510 17:08:56.573729 10926 solver.cpp:406] Test net output #104: loss3/accuracy06 = 0.425 | |
| I0510 17:08:56.573741 10926 solver.cpp:406] Test net output #105: loss3/accuracy07 = 0.677 | |
| I0510 17:08:56.573753 10926 solver.cpp:406] Test net output #106: loss3/accuracy08 = 0.819 | |
| I0510 17:08:56.573765 10926 solver.cpp:406] Test net output #107: loss3/accuracy09 = 0.902 | |
| I0510 17:08:56.573776 10926 solver.cpp:406] Test net output #108: loss3/accuracy10 = 0.919 | |
| I0510 17:08:56.573788 10926 solver.cpp:406] Test net output #109: loss3/accuracy11 = 0.936 | |
| I0510 17:08:56.573799 10926 solver.cpp:406] Test net output #110: loss3/accuracy12 = 0.946 | |
| I0510 17:08:56.573812 10926 solver.cpp:406] Test net output #111: loss3/accuracy13 = 0.961 | |
| I0510 17:08:56.573823 10926 solver.cpp:406] Test net output #112: loss3/accuracy14 = 0.971 | |
| I0510 17:08:56.573835 10926 solver.cpp:406] Test net output #113: loss3/accuracy15 = 0.978 | |
| I0510 17:08:56.573846 10926 solver.cpp:406] Test net output #114: loss3/accuracy16 = 0.984 | |
| I0510 17:08:56.573868 10926 solver.cpp:406] Test net output #115: loss3/accuracy17 = 0.994 | |
| I0510 17:08:56.573881 10926 solver.cpp:406] Test net output #116: loss3/accuracy18 = 0.996 | |
| I0510 17:08:56.573894 10926 solver.cpp:406] Test net output #117: loss3/accuracy19 = 0.997 | |
| I0510 17:08:56.573904 10926 solver.cpp:406] Test net output #118: loss3/accuracy20 = 0.997 | |
| I0510 17:08:56.573917 10926 solver.cpp:406] Test net output #119: loss3/accuracy21 = 0.998 | |
| I0510 17:08:56.573931 10926 solver.cpp:406] Test net output #120: loss3/accuracy22 = 0.999 | |
| I0510 17:08:56.573945 10926 solver.cpp:406] Test net output #121: loss3/accuracy_incl_empty = 0.738545 | |
| I0510 17:08:56.573956 10926 solver.cpp:406] Test net output #122: loss3/accuracy_top3 = 0.268879 | |
| I0510 17:08:56.573971 10926 solver.cpp:406] Test net output #123: loss3/cross_entropy_loss = 3.05473 (* 1 = 3.05473 loss) | |
| I0510 17:08:56.573983 10926 solver.cpp:406] Test net output #124: loss3/cross_entropy_loss_incl_empty = 0.94447 (* 1 = 0.94447 loss) | |
| I0510 17:08:56.573997 10926 solver.cpp:406] Test net output #125: loss3/loss01 = 2.80309 (* 0.0909091 = 0.254827 loss) | |
| I0510 17:08:56.574012 10926 solver.cpp:406] Test net output #126: loss3/loss02 = 2.98805 (* 0.0909091 = 0.271641 loss) | |
| I0510 17:08:56.574025 10926 solver.cpp:406] Test net output #127: loss3/loss03 = 3.05628 (* 0.0909091 = 0.277844 loss) | |
| I0510 17:08:56.574038 10926 solver.cpp:406] Test net output #128: loss3/loss04 = 2.97113 (* 0.0909091 = 0.270103 loss) | |
| I0510 17:08:56.574051 10926 solver.cpp:406] Test net output #129: loss3/loss05 = 2.55388 (* 0.0909091 = 0.232171 loss) | |
| I0510 17:08:56.574065 10926 solver.cpp:406] Test net output #130: loss3/loss06 = 2.16229 (* 0.0909091 = 0.196572 loss) | |
| I0510 17:08:56.574079 10926 solver.cpp:406] Test net output #131: loss3/loss07 = 1.37835 (* 0.0909091 = 0.125304 loss) | |
| I0510 17:08:56.574095 10926 solver.cpp:406] Test net output #132: loss3/loss08 = 0.777529 (* 0.0909091 = 0.0706845 loss) | |
| I0510 17:08:56.574110 10926 solver.cpp:406] Test net output #133: loss3/loss09 = 0.410001 (* 0.0909091 = 0.0372729 loss) | |
| I0510 17:08:56.574123 10926 solver.cpp:406] Test net output #134: loss3/loss10 = 0.341959 (* 0.0909091 = 0.0310872 loss) | |
| I0510 17:08:56.574137 10926 solver.cpp:406] Test net output #135: loss3/loss11 = 0.272848 (* 0.0909091 = 0.0248043 loss) | |
| I0510 17:08:56.574151 10926 solver.cpp:406] Test net output #136: loss3/loss12 = 0.24906 (* 0.0909091 = 0.0226418 loss) | |
| I0510 17:08:56.574164 10926 solver.cpp:406] Test net output #137: loss3/loss13 = 0.187615 (* 0.0909091 = 0.0170559 loss) | |
| I0510 17:08:56.574178 10926 solver.cpp:406] Test net output #138: loss3/loss14 = 0.139864 (* 0.0909091 = 0.0127149 loss) | |
| I0510 17:08:56.574193 10926 solver.cpp:406] Test net output #139: loss3/loss15 = 0.105012 (* 0.0909091 = 0.00954657 loss) | |
| I0510 17:08:56.574206 10926 solver.cpp:406] Test net output #140: loss3/loss16 = 0.089348 (* 0.0909091 = 0.00812255 loss) | |
| I0510 17:08:56.574219 10926 solver.cpp:406] Test net output #141: loss3/loss17 = 0.0402052 (* 0.0909091 = 0.00365502 loss) | |
| I0510 17:08:56.574234 10926 solver.cpp:406] Test net output #142: loss3/loss18 = 0.0298108 (* 0.0909091 = 0.00271007 loss) | |
| I0510 17:08:56.574246 10926 solver.cpp:406] Test net output #143: loss3/loss19 = 0.0236894 (* 0.0909091 = 0.00215358 loss) | |
| I0510 17:08:56.574260 10926 solver.cpp:406] Test net output #144: loss3/loss20 = 0.0273213 (* 0.0909091 = 0.00248375 loss) | |
| I0510 17:08:56.574275 10926 solver.cpp:406] Test net output #145: loss3/loss21 = 0.0172842 (* 0.0909091 = 0.00157129 loss) | |
| I0510 17:08:56.574288 10926 solver.cpp:406] Test net output #146: loss3/loss22 = 0.0109751 (* 0.0909091 = 0.000997736 loss) | |
| I0510 17:08:56.574301 10926 solver.cpp:406] Test net output #147: total_accuracy = 0 | |
| I0510 17:08:56.574311 10926 solver.cpp:406] Test net output #148: total_accuracy_not_rec = 0 | |
| I0510 17:08:56.574322 10926 solver.cpp:406] Test net output #149: total_confidence = 9.3692e-05 | |
| I0510 17:08:56.574345 10926 solver.cpp:406] Test net output #150: total_confidence_not_rec = 0.000128046 | |
| I0510 17:08:56.720222 10926 solver.cpp:229] Iteration 20000, loss = 9.68413 | |
| I0510 17:08:56.720289 10926 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0833333 | |
| I0510 17:08:56.720309 10926 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0510 17:08:56.720322 10926 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.125 | |
| I0510 17:08:56.720335 10926 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.25 | |
| I0510 17:08:56.720348 10926 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.625 | |
| I0510 17:08:56.720361 10926 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.5 | |
| I0510 17:08:56.720373 10926 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.75 | |
| I0510 17:08:56.720386 10926 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.875 | |
| I0510 17:08:56.720399 10926 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.875 | |
| I0510 17:08:56.720412 10926 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0510 17:08:56.720424 10926 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0510 17:08:56.720440 10926 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0510 17:08:56.720453 10926 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0510 17:08:56.720465 10926 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0510 17:08:56.720477 10926 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0510 17:08:56.720489 10926 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0510 17:08:56.720501 10926 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0510 17:08:56.720513 10926 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0510 17:08:56.720525 10926 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0510 17:08:56.720536 10926 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0510 17:08:56.720548 10926 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0510 17:08:56.720561 10926 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0510 17:08:56.720572 10926 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0510 17:08:56.720584 10926 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.806818 | |
| I0510 17:08:56.720597 10926 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.25 | |
| I0510 17:08:56.720613 10926 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 2.97426 (* 0.3 = 0.892278 loss) | |
| I0510 17:08:56.720628 10926 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 0.70382 (* 0.3 = 0.211146 loss) | |
| I0510 17:08:56.720643 10926 solver.cpp:245] Train net output #27: loss1/loss01 = 3.11188 (* 0.0272727 = 0.0848695 loss) | |
| I0510 17:08:56.720659 10926 solver.cpp:245] Train net output #28: loss1/loss02 = 2.93765 (* 0.0272727 = 0.0801178 loss) | |
| I0510 17:08:56.720674 10926 solver.cpp:245] Train net output #29: loss1/loss03 = 3.02539 (* 0.0272727 = 0.0825107 loss) | |
| I0510 17:08:56.720688 10926 solver.cpp:245] Train net output #30: loss1/loss04 = 1.85847 (* 0.0272727 = 0.0506856 loss) | |
| I0510 17:08:56.720702 10926 solver.cpp:245] Train net output #31: loss1/loss05 = 1.888 (* 0.0272727 = 0.0514908 loss) | |
| I0510 17:08:56.720717 10926 solver.cpp:245] Train net output #32: loss1/loss06 = 1.46238 (* 0.0272727 = 0.0398831 loss) | |
| I0510 17:08:56.720731 10926 solver.cpp:245] Train net output #33: loss1/loss07 = 0.560448 (* 0.0272727 = 0.0152849 loss) | |
| I0510 17:08:56.720746 10926 solver.cpp:245] Train net output #34: loss1/loss08 = 0.535915 (* 0.0272727 = 0.0146159 loss) | |
| I0510 17:08:56.720762 10926 solver.cpp:245] Train net output #35: loss1/loss09 = 0.0224845 (* 0.0272727 = 0.000613213 loss) | |
| I0510 17:08:56.720777 10926 solver.cpp:245] Train net output #36: loss1/loss10 = 0.0293871 (* 0.0272727 = 0.000801467 loss) | |
| I0510 17:08:56.720790 10926 solver.cpp:245] Train net output #37: loss1/loss11 = 0.0165188 (* 0.0272727 = 0.000450513 loss) | |
| I0510 17:08:56.720832 10926 solver.cpp:245] Train net output #38: loss1/loss12 = 0.0200689 (* 0.0272727 = 0.000547332 loss) | |
| I0510 17:08:56.720849 10926 solver.cpp:245] Train net output #39: loss1/loss13 = 0.00470049 (* 0.0272727 = 0.000128195 loss) | |
| I0510 17:08:56.720862 10926 solver.cpp:245] Train net output #40: loss1/loss14 = 0.010526 (* 0.0272727 = 0.000287073 loss) | |
| I0510 17:08:56.720877 10926 solver.cpp:245] Train net output #41: loss1/loss15 = 0.00551302 (* 0.0272727 = 0.000150355 loss) | |
| I0510 17:08:56.720892 10926 solver.cpp:245] Train net output #42: loss1/loss16 = 0.00493167 (* 0.0272727 = 0.0001345 loss) | |
| I0510 17:08:56.720907 10926 solver.cpp:245] Train net output #43: loss1/loss17 = 0.00343279 (* 0.0272727 = 9.36217e-05 loss) | |
| I0510 17:08:56.720921 10926 solver.cpp:245] Train net output #44: loss1/loss18 = 0.00316587 (* 0.0272727 = 8.63419e-05 loss) | |
| I0510 17:08:56.720935 10926 solver.cpp:245] Train net output #45: loss1/loss19 = 0.00436962 (* 0.0272727 = 0.000119171 loss) | |
| I0510 17:08:56.720950 10926 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00248001 (* 0.0272727 = 6.76367e-05 loss) | |
| I0510 17:08:56.720964 10926 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00277478 (* 0.0272727 = 7.56758e-05 loss) | |
| I0510 17:08:56.720979 10926 solver.cpp:245] Train net output #48: loss1/loss22 = 0.00177259 (* 0.0272727 = 4.83435e-05 loss) | |
| I0510 17:08:56.720991 10926 solver.cpp:245] Train net output #49: loss2/accuracy = 0.111111 | |
| I0510 17:08:56.721004 10926 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.125 | |
| I0510 17:08:56.721017 10926 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0510 17:08:56.721029 10926 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0.125 | |
| I0510 17:08:56.721041 10926 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.625 | |
| I0510 17:08:56.721053 10926 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.625 | |
| I0510 17:08:56.721065 10926 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.625 | |
| I0510 17:08:56.721077 10926 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.875 | |
| I0510 17:08:56.721089 10926 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.875 | |
| I0510 17:08:56.721101 10926 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0510 17:08:56.721113 10926 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0510 17:08:56.721140 10926 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0510 17:08:56.721154 10926 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0510 17:08:56.721166 10926 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0510 17:08:56.721179 10926 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0510 17:08:56.721190 10926 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0510 17:08:56.721202 10926 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0510 17:08:56.721213 10926 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0510 17:08:56.721225 10926 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0510 17:08:56.721237 10926 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0510 17:08:56.721249 10926 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0510 17:08:56.721261 10926 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0510 17:08:56.721272 10926 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0510 17:08:56.721284 10926 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.795455 | |
| I0510 17:08:56.721297 10926 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.333333 | |
| I0510 17:08:56.721310 10926 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 2.92058 (* 0.3 = 0.876175 loss) | |
| I0510 17:08:56.721325 10926 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 0.776029 (* 0.3 = 0.232809 loss) | |
| I0510 17:08:56.721354 10926 solver.cpp:245] Train net output #76: loss2/loss01 = 3.07569 (* 0.0272727 = 0.0838825 loss) | |
| I0510 17:08:56.721369 10926 solver.cpp:245] Train net output #77: loss2/loss02 = 3.04056 (* 0.0272727 = 0.0829243 loss) | |
| I0510 17:08:56.721384 10926 solver.cpp:245] Train net output #78: loss2/loss03 = 3.32354 (* 0.0272727 = 0.090642 loss) | |
| I0510 17:08:56.721398 10926 solver.cpp:245] Train net output #79: loss2/loss04 = 2.07576 (* 0.0272727 = 0.0566116 loss) | |
| I0510 17:08:56.721412 10926 solver.cpp:245] Train net output #80: loss2/loss05 = 1.53229 (* 0.0272727 = 0.0417899 loss) | |
| I0510 17:08:56.721426 10926 solver.cpp:245] Train net output #81: loss2/loss06 = 1.6342 (* 0.0272727 = 0.0445691 loss) | |
| I0510 17:08:56.721441 10926 solver.cpp:245] Train net output #82: loss2/loss07 = 0.518082 (* 0.0272727 = 0.0141295 loss) | |
| I0510 17:08:56.721454 10926 solver.cpp:245] Train net output #83: loss2/loss08 = 0.454933 (* 0.0272727 = 0.0124073 loss) | |
| I0510 17:08:56.721469 10926 solver.cpp:245] Train net output #84: loss2/loss09 = 0.00955323 (* 0.0272727 = 0.000260543 loss) | |
| I0510 17:08:56.721488 10926 solver.cpp:245] Train net output #85: loss2/loss10 = 0.00610217 (* 0.0272727 = 0.000166423 loss) | |
| I0510 17:08:56.721503 10926 solver.cpp:245] Train net output #86: loss2/loss11 = 0.00407054 (* 0.0272727 = 0.000111015 loss) | |
| I0510 17:08:56.721518 10926 solver.cpp:245] Train net output #87: loss2/loss12 = 0.00201647 (* 0.0272727 = 5.49946e-05 loss) | |
| I0510 17:08:56.721531 10926 solver.cpp:245] Train net output #88: loss2/loss13 = 0.00365455 (* 0.0272727 = 9.96696e-05 loss) | |
| I0510 17:08:56.721546 10926 solver.cpp:245] Train net output #89: loss2/loss14 = 0.00207833 (* 0.0272727 = 5.66818e-05 loss) | |
| I0510 17:08:56.721560 10926 solver.cpp:245] Train net output #90: loss2/loss15 = 0.00218411 (* 0.0272727 = 5.95666e-05 loss) | |
| I0510 17:08:56.721575 10926 solver.cpp:245] Train net output #91: loss2/loss16 = 0.00169749 (* 0.0272727 = 4.6295e-05 loss) | |
| I0510 17:08:56.721588 10926 solver.cpp:245] Train net output #92: loss2/loss17 = 0.00240131 (* 0.0272727 = 6.54902e-05 loss) | |
| I0510 17:08:56.721603 10926 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0015682 (* 0.0272727 = 4.2769e-05 loss) | |
| I0510 17:08:56.721617 10926 solver.cpp:245] Train net output #94: loss2/loss19 = 0.00159604 (* 0.0272727 = 4.35283e-05 loss) | |
| I0510 17:08:56.721632 10926 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00151866 (* 0.0272727 = 4.14181e-05 loss) | |
| I0510 17:08:56.721647 10926 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00119738 (* 0.0272727 = 3.26558e-05 loss) | |
| I0510 17:08:56.721660 10926 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00155028 (* 0.0272727 = 4.22804e-05 loss) | |
| I0510 17:08:56.721673 10926 solver.cpp:245] Train net output #98: loss3/accuracy = 0.138889 | |
| I0510 17:08:56.721685 10926 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.125 | |
| I0510 17:08:56.721699 10926 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0510 17:08:56.721712 10926 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0.125 | |
| I0510 17:08:56.721724 10926 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.625 | |
| I0510 17:08:56.721735 10926 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.625 | |
| I0510 17:08:56.721747 10926 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.75 | |
| I0510 17:08:56.721760 10926 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.875 | |
| I0510 17:08:56.721771 10926 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.875 | |
| I0510 17:08:56.721783 10926 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0510 17:08:56.721794 10926 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0510 17:08:56.721807 10926 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0510 17:08:56.721818 10926 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0510 17:08:56.721829 10926 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0510 17:08:56.721853 10926 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0510 17:08:56.721866 10926 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0510 17:08:56.721879 10926 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0510 17:08:56.721890 10926 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0510 17:08:56.721902 10926 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0510 17:08:56.721914 10926 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0510 17:08:56.721925 10926 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0510 17:08:56.721936 10926 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0510 17:08:56.721948 10926 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0510 17:08:56.721961 10926 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.795455 | |
| I0510 17:08:56.721972 10926 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.361111 | |
| I0510 17:08:56.721987 10926 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 2.80601 (* 1 = 2.80601 loss) | |
| I0510 17:08:56.722000 10926 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 0.719447 (* 1 = 0.719447 loss) | |
| I0510 17:08:56.722015 10926 solver.cpp:245] Train net output #125: loss3/loss01 = 2.55988 (* 0.0909091 = 0.232716 loss) | |
| I0510 17:08:56.722029 10926 solver.cpp:245] Train net output #126: loss3/loss02 = 2.51829 (* 0.0909091 = 0.228935 loss) | |
| I0510 17:08:56.722043 10926 solver.cpp:245] Train net output #127: loss3/loss03 = 3.02107 (* 0.0909091 = 0.274642 loss) | |
| I0510 17:08:56.722057 10926 solver.cpp:245] Train net output #128: loss3/loss04 = 1.96598 (* 0.0909091 = 0.178726 loss) | |
| I0510 17:08:56.722071 10926 solver.cpp:245] Train net output #129: loss3/loss05 = 1.35856 (* 0.0909091 = 0.123505 loss) | |
| I0510 17:08:56.722085 10926 solver.cpp:245] Train net output #130: loss3/loss06 = 1.22775 (* 0.0909091 = 0.111614 loss) | |
| I0510 17:08:56.722100 10926 solver.cpp:245] Train net output #131: loss3/loss07 = 0.564231 (* 0.0909091 = 0.0512938 loss) | |
| I0510 17:08:56.722113 10926 solver.cpp:245] Train net output #132: loss3/loss08 = 0.34575 (* 0.0909091 = 0.0314318 loss) | |
| I0510 17:08:56.722128 10926 solver.cpp:245] Train net output #133: loss3/loss09 = 0.00703671 (* 0.0909091 = 0.000639701 loss) | |
| I0510 17:08:56.722142 10926 solver.cpp:245] Train net output #134: loss3/loss10 = 0.00327888 (* 0.0909091 = 0.00029808 loss) | |
| I0510 17:08:56.722157 10926 solver.cpp:245] Train net output #135: loss3/loss11 = 0.00214069 (* 0.0909091 = 0.000194608 loss) | |
| I0510 17:08:56.722170 10926 solver.cpp:245] Train net output #136: loss3/loss12 = 0.00128646 (* 0.0909091 = 0.000116951 loss) | |
| I0510 17:08:56.722185 10926 solver.cpp:245] Train net output #137: loss3/loss13 = 0.00145627 (* 0.0909091 = 0.000132388 loss) | |
| I0510 17:08:56.722199 10926 solver.cpp:245] Train net output #138: loss3/loss14 = 0.00109318 (* 0.0909091 = 9.93803e-05 loss) | |
| I0510 17:08:56.722213 10926 solver.cpp:245] Train net output #139: loss3/loss15 = 0.000566114 (* 0.0909091 = 5.14649e-05 loss) | |
| I0510 17:08:56.722228 10926 solver.cpp:245] Train net output #140: loss3/loss16 = 0.000743929 (* 0.0909091 = 6.76299e-05 loss) | |
| I0510 17:08:56.722242 10926 solver.cpp:245] Train net output #141: loss3/loss17 = 0.000600585 (* 0.0909091 = 5.45987e-05 loss) | |
| I0510 17:08:56.722256 10926 solver.cpp:245] Train net output #142: loss3/loss18 = 0.000304625 (* 0.0909091 = 2.76932e-05 loss) | |
| I0510 17:08:56.722271 10926 solver.cpp:245] Train net output #143: loss3/loss19 = 0.000283003 (* 0.0909091 = 2.57276e-05 loss) | |
| I0510 17:08:56.722285 10926 solver.cpp:245] Train net output #144: loss3/loss20 = 0.000247811 (* 0.0909091 = 2.25283e-05 loss) | |
| I0510 17:08:56.722300 10926 solver.cpp:245] Train net output #145: loss3/loss21 = 0.000262051 (* 0.0909091 = 2.38228e-05 loss) | |
| I0510 17:08:56.722313 10926 solver.cpp:245] Train net output #146: loss3/loss22 = 0.000175025 (* 0.0909091 = 1.59114e-05 loss) | |
| I0510 17:08:56.722337 10926 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0510 17:08:56.722350 10926 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0510 17:08:56.722362 10926 solver.cpp:245] Train net output #149: total_confidence = 0.000133859 | |
| I0510 17:08:56.722373 10926 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.000696146 | |
| I0510 17:08:56.722388 10926 sgd_solver.cpp:106] Iteration 20000, lr = 0.001 | |
| I0510 17:09:26.555831 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 38.8658 > 30) by scale factor 0.771887 | |
| I0510 17:09:34.204459 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 45.0792 > 30) by scale factor 0.665496 | |
| I0510 17:10:46.537307 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 44.316 > 30) by scale factor 0.676956 | |
| I0510 17:10:55.623013 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 32.6152 > 30) by scale factor 0.919818 | |
| I0510 17:11:23.733506 10926 solver.cpp:229] Iteration 20500, loss = 9.71722 | |
| I0510 17:11:23.733641 10926 solver.cpp:245] Train net output #0: loss1/accuracy = 0.133333 | |
| I0510 17:11:23.733664 10926 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.25 | |
| I0510 17:11:23.733677 10926 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0510 17:11:23.733691 10926 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.125 | |
| I0510 17:11:23.733703 10926 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.125 | |
| I0510 17:11:23.733716 10926 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.25 | |
| I0510 17:11:23.733729 10926 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.25 | |
| I0510 17:11:23.733742 10926 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.5 | |
| I0510 17:11:23.733755 10926 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.875 | |
| I0510 17:11:23.733768 10926 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.75 | |
| I0510 17:11:23.733782 10926 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0510 17:11:23.733794 10926 solver.cpp:245] Train net output #11: loss1/accuracy11 = 0.875 | |
| I0510 17:11:23.733806 10926 solver.cpp:245] Train net output #12: loss1/accuracy12 = 0.875 | |
| I0510 17:11:23.733820 10926 solver.cpp:245] Train net output #13: loss1/accuracy13 = 0.875 | |
| I0510 17:11:23.733832 10926 solver.cpp:245] Train net output #14: loss1/accuracy14 = 0.875 | |
| I0510 17:11:23.733844 10926 solver.cpp:245] Train net output #15: loss1/accuracy15 = 0.875 | |
| I0510 17:11:23.733857 10926 solver.cpp:245] Train net output #16: loss1/accuracy16 = 0.875 | |
| I0510 17:11:23.733870 10926 solver.cpp:245] Train net output #17: loss1/accuracy17 = 0.875 | |
| I0510 17:11:23.733885 10926 solver.cpp:245] Train net output #18: loss1/accuracy18 = 0.875 | |
| I0510 17:11:23.733897 10926 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0510 17:11:23.733909 10926 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0510 17:11:23.733922 10926 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0510 17:11:23.733934 10926 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0510 17:11:23.733947 10926 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.6875 | |
| I0510 17:11:23.733958 10926 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.3 | |
| I0510 17:11:23.733975 10926 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.05286 (* 0.3 = 0.915857 loss) | |
| I0510 17:11:23.733990 10926 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.22894 (* 0.3 = 0.368681 loss) | |
| I0510 17:11:23.734005 10926 solver.cpp:245] Train net output #27: loss1/loss01 = 3.01764 (* 0.0272727 = 0.0822992 loss) | |
| I0510 17:11:23.734019 10926 solver.cpp:245] Train net output #28: loss1/loss02 = 2.75286 (* 0.0272727 = 0.075078 loss) | |
| I0510 17:11:23.734033 10926 solver.cpp:245] Train net output #29: loss1/loss03 = 3.95712 (* 0.0272727 = 0.107922 loss) | |
| I0510 17:11:23.734048 10926 solver.cpp:245] Train net output #30: loss1/loss04 = 3.4709 (* 0.0272727 = 0.0946608 loss) | |
| I0510 17:11:23.734062 10926 solver.cpp:245] Train net output #31: loss1/loss05 = 2.64748 (* 0.0272727 = 0.072204 loss) | |
| I0510 17:11:23.734076 10926 solver.cpp:245] Train net output #32: loss1/loss06 = 2.45432 (* 0.0272727 = 0.066936 loss) | |
| I0510 17:11:23.734091 10926 solver.cpp:245] Train net output #33: loss1/loss07 = 2.06442 (* 0.0272727 = 0.0563023 loss) | |
| I0510 17:11:23.734104 10926 solver.cpp:245] Train net output #34: loss1/loss08 = 0.659908 (* 0.0272727 = 0.0179975 loss) | |
| I0510 17:11:23.734119 10926 solver.cpp:245] Train net output #35: loss1/loss09 = 1.18975 (* 0.0272727 = 0.0324476 loss) | |
| I0510 17:11:23.734133 10926 solver.cpp:245] Train net output #36: loss1/loss10 = 0.43682 (* 0.0272727 = 0.0119133 loss) | |
| I0510 17:11:23.734148 10926 solver.cpp:245] Train net output #37: loss1/loss11 = 0.54702 (* 0.0272727 = 0.0149187 loss) | |
| I0510 17:11:23.734163 10926 solver.cpp:245] Train net output #38: loss1/loss12 = 0.533351 (* 0.0272727 = 0.0145459 loss) | |
| I0510 17:11:23.734196 10926 solver.cpp:245] Train net output #39: loss1/loss13 = 0.406922 (* 0.0272727 = 0.0110979 loss) | |
| I0510 17:11:23.734213 10926 solver.cpp:245] Train net output #40: loss1/loss14 = 0.708222 (* 0.0272727 = 0.0193152 loss) | |
| I0510 17:11:23.734227 10926 solver.cpp:245] Train net output #41: loss1/loss15 = 0.508259 (* 0.0272727 = 0.0138616 loss) | |
| I0510 17:11:23.734242 10926 solver.cpp:245] Train net output #42: loss1/loss16 = 0.566849 (* 0.0272727 = 0.0154595 loss) | |
| I0510 17:11:23.734256 10926 solver.cpp:245] Train net output #43: loss1/loss17 = 0.684821 (* 0.0272727 = 0.0186769 loss) | |
| I0510 17:11:23.734272 10926 solver.cpp:245] Train net output #44: loss1/loss18 = 0.985354 (* 0.0272727 = 0.0268733 loss) | |
| I0510 17:11:23.734285 10926 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0213479 (* 0.0272727 = 0.000582216 loss) | |
| I0510 17:11:23.734300 10926 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00872944 (* 0.0272727 = 0.000238076 loss) | |
| I0510 17:11:23.734314 10926 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00514885 (* 0.0272727 = 0.000140423 loss) | |
| I0510 17:11:23.734329 10926 solver.cpp:245] Train net output #48: loss1/loss22 = 0.00578225 (* 0.0272727 = 0.000157698 loss) | |
| I0510 17:11:23.734341 10926 solver.cpp:245] Train net output #49: loss2/accuracy = 0.1 | |
| I0510 17:11:23.734354 10926 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.25 | |
| I0510 17:11:23.734366 10926 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0510 17:11:23.734380 10926 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0510 17:11:23.734391 10926 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.25 | |
| I0510 17:11:23.734403 10926 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.25 | |
| I0510 17:11:23.734416 10926 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.25 | |
| I0510 17:11:23.734427 10926 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.625 | |
| I0510 17:11:23.734439 10926 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.875 | |
| I0510 17:11:23.734452 10926 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.75 | |
| I0510 17:11:23.734464 10926 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.875 | |
| I0510 17:11:23.734477 10926 solver.cpp:245] Train net output #60: loss2/accuracy11 = 0.875 | |
| I0510 17:11:23.734488 10926 solver.cpp:245] Train net output #61: loss2/accuracy12 = 0.875 | |
| I0510 17:11:23.734500 10926 solver.cpp:245] Train net output #62: loss2/accuracy13 = 0.875 | |
| I0510 17:11:23.734513 10926 solver.cpp:245] Train net output #63: loss2/accuracy14 = 0.875 | |
| I0510 17:11:23.734525 10926 solver.cpp:245] Train net output #64: loss2/accuracy15 = 0.875 | |
| I0510 17:11:23.734537 10926 solver.cpp:245] Train net output #65: loss2/accuracy16 = 0.875 | |
| I0510 17:11:23.734549 10926 solver.cpp:245] Train net output #66: loss2/accuracy17 = 0.875 | |
| I0510 17:11:23.734561 10926 solver.cpp:245] Train net output #67: loss2/accuracy18 = 0.875 | |
| I0510 17:11:23.734575 10926 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0510 17:11:23.734586 10926 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0510 17:11:23.734598 10926 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0510 17:11:23.734611 10926 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0510 17:11:23.734622 10926 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.6875 | |
| I0510 17:11:23.734634 10926 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.316667 | |
| I0510 17:11:23.734648 10926 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.0657 (* 0.3 = 0.919711 loss) | |
| I0510 17:11:23.734666 10926 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.16858 (* 0.3 = 0.350575 loss) | |
| I0510 17:11:23.734681 10926 solver.cpp:245] Train net output #76: loss2/loss01 = 2.93233 (* 0.0272727 = 0.0799726 loss) | |
| I0510 17:11:23.734695 10926 solver.cpp:245] Train net output #77: loss2/loss02 = 3.23384 (* 0.0272727 = 0.0881957 loss) | |
| I0510 17:11:23.734722 10926 solver.cpp:245] Train net output #78: loss2/loss03 = 3.66014 (* 0.0272727 = 0.0998219 loss) | |
| I0510 17:11:23.734737 10926 solver.cpp:245] Train net output #79: loss2/loss04 = 3.42273 (* 0.0272727 = 0.0933471 loss) | |
| I0510 17:11:23.734752 10926 solver.cpp:245] Train net output #80: loss2/loss05 = 2.88339 (* 0.0272727 = 0.0786379 loss) | |
| I0510 17:11:23.734766 10926 solver.cpp:245] Train net output #81: loss2/loss06 = 2.71602 (* 0.0272727 = 0.0740732 loss) | |
| I0510 17:11:23.734781 10926 solver.cpp:245] Train net output #82: loss2/loss07 = 1.80212 (* 0.0272727 = 0.0491487 loss) | |
| I0510 17:11:23.734796 10926 solver.cpp:245] Train net output #83: loss2/loss08 = 0.625008 (* 0.0272727 = 0.0170457 loss) | |
| I0510 17:11:23.734809 10926 solver.cpp:245] Train net output #84: loss2/loss09 = 1.21811 (* 0.0272727 = 0.0332213 loss) | |
| I0510 17:11:23.734823 10926 solver.cpp:245] Train net output #85: loss2/loss10 = 0.478461 (* 0.0272727 = 0.0130489 loss) | |
| I0510 17:11:23.734838 10926 solver.cpp:245] Train net output #86: loss2/loss11 = 0.536962 (* 0.0272727 = 0.0146444 loss) | |
| I0510 17:11:23.734851 10926 solver.cpp:245] Train net output #87: loss2/loss12 = 0.36989 (* 0.0272727 = 0.0100879 loss) | |
| I0510 17:11:23.734865 10926 solver.cpp:245] Train net output #88: loss2/loss13 = 0.487538 (* 0.0272727 = 0.0132965 loss) | |
| I0510 17:11:23.734879 10926 solver.cpp:245] Train net output #89: loss2/loss14 = 0.473178 (* 0.0272727 = 0.0129049 loss) | |
| I0510 17:11:23.734894 10926 solver.cpp:245] Train net output #90: loss2/loss15 = 0.421388 (* 0.0272727 = 0.0114924 loss) | |
| I0510 17:11:23.734907 10926 solver.cpp:245] Train net output #91: loss2/loss16 = 0.569074 (* 0.0272727 = 0.0155202 loss) | |
| I0510 17:11:23.734921 10926 solver.cpp:245] Train net output #92: loss2/loss17 = 0.798424 (* 0.0272727 = 0.0217752 loss) | |
| I0510 17:11:23.734938 10926 solver.cpp:245] Train net output #93: loss2/loss18 = 0.92414 (* 0.0272727 = 0.0252038 loss) | |
| I0510 17:11:23.734953 10926 solver.cpp:245] Train net output #94: loss2/loss19 = 0.011819 (* 0.0272727 = 0.000322336 loss) | |
| I0510 17:11:23.734968 10926 solver.cpp:245] Train net output #95: loss2/loss20 = 0.0139827 (* 0.0272727 = 0.000381346 loss) | |
| I0510 17:11:23.734982 10926 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00765714 (* 0.0272727 = 0.000208831 loss) | |
| I0510 17:11:23.734997 10926 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00516336 (* 0.0272727 = 0.000140819 loss) | |
| I0510 17:11:23.735009 10926 solver.cpp:245] Train net output #98: loss3/accuracy = 0.1 | |
| I0510 17:11:23.735021 10926 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.125 | |
| I0510 17:11:23.735034 10926 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.125 | |
| I0510 17:11:23.735046 10926 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0.125 | |
| I0510 17:11:23.735059 10926 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0510 17:11:23.735070 10926 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.25 | |
| I0510 17:11:23.735082 10926 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.25 | |
| I0510 17:11:23.735095 10926 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.5 | |
| I0510 17:11:23.735106 10926 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.875 | |
| I0510 17:11:23.735118 10926 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.75 | |
| I0510 17:11:23.735131 10926 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.875 | |
| I0510 17:11:23.735142 10926 solver.cpp:245] Train net output #109: loss3/accuracy11 = 0.875 | |
| I0510 17:11:23.735154 10926 solver.cpp:245] Train net output #110: loss3/accuracy12 = 0.875 | |
| I0510 17:11:23.735167 10926 solver.cpp:245] Train net output #111: loss3/accuracy13 = 0.875 | |
| I0510 17:11:23.735178 10926 solver.cpp:245] Train net output #112: loss3/accuracy14 = 0.875 | |
| I0510 17:11:23.735189 10926 solver.cpp:245] Train net output #113: loss3/accuracy15 = 0.875 | |
| I0510 17:11:23.735201 10926 solver.cpp:245] Train net output #114: loss3/accuracy16 = 0.875 | |
| I0510 17:11:23.735224 10926 solver.cpp:245] Train net output #115: loss3/accuracy17 = 0.875 | |
| I0510 17:11:23.735239 10926 solver.cpp:245] Train net output #116: loss3/accuracy18 = 0.875 | |
| I0510 17:11:23.735250 10926 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0510 17:11:23.735262 10926 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0510 17:11:23.735275 10926 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0510 17:11:23.735286 10926 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0510 17:11:23.735298 10926 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.6875 | |
| I0510 17:11:23.735311 10926 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.283333 | |
| I0510 17:11:23.735326 10926 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 2.92552 (* 1 = 2.92552 loss) | |
| I0510 17:11:23.735339 10926 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.09527 (* 1 = 1.09527 loss) | |
| I0510 17:11:23.735353 10926 solver.cpp:245] Train net output #125: loss3/loss01 = 2.58389 (* 0.0909091 = 0.234899 loss) | |
| I0510 17:11:23.735368 10926 solver.cpp:245] Train net output #126: loss3/loss02 = 2.42387 (* 0.0909091 = 0.220352 loss) | |
| I0510 17:11:23.735381 10926 solver.cpp:245] Train net output #127: loss3/loss03 = 3.47235 (* 0.0909091 = 0.315668 loss) | |
| I0510 17:11:23.735395 10926 solver.cpp:245] Train net output #128: loss3/loss04 = 3.12613 (* 0.0909091 = 0.284194 loss) | |
| I0510 17:11:23.735410 10926 solver.cpp:245] Train net output #129: loss3/loss05 = 2.46222 (* 0.0909091 = 0.223838 loss) | |
| I0510 17:11:23.735424 10926 solver.cpp:245] Train net output #130: loss3/loss06 = 2.43258 (* 0.0909091 = 0.221144 loss) | |
| I0510 17:11:23.735442 10926 solver.cpp:245] Train net output #131: loss3/loss07 = 1.73324 (* 0.0909091 = 0.157567 loss) | |
| I0510 17:11:23.735452 10926 solver.cpp:245] Train net output #132: loss3/loss08 = 0.672987 (* 0.0909091 = 0.0611807 loss) | |
| I0510 17:11:23.735467 10926 solver.cpp:245] Train net output #133: loss3/loss09 = 0.979811 (* 0.0909091 = 0.0890737 loss) | |
| I0510 17:11:23.735481 10926 solver.cpp:245] Train net output #134: loss3/loss10 = 0.427515 (* 0.0909091 = 0.038865 loss) | |
| I0510 17:11:23.735496 10926 solver.cpp:245] Train net output #135: loss3/loss11 = 0.383601 (* 0.0909091 = 0.0348728 loss) | |
| I0510 17:11:23.735509 10926 solver.cpp:245] Train net output #136: loss3/loss12 = 0.365065 (* 0.0909091 = 0.0331878 loss) | |
| I0510 17:11:23.735524 10926 solver.cpp:245] Train net output #137: loss3/loss13 = 0.400414 (* 0.0909091 = 0.0364013 loss) | |
| I0510 17:11:23.735538 10926 solver.cpp:245] Train net output #138: loss3/loss14 = 0.57573 (* 0.0909091 = 0.0523391 loss) | |
| I0510 17:11:23.735551 10926 solver.cpp:245] Train net output #139: loss3/loss15 = 0.459799 (* 0.0909091 = 0.0417999 loss) | |
| I0510 17:11:23.735565 10926 solver.cpp:245] Train net output #140: loss3/loss16 = 0.445025 (* 0.0909091 = 0.0404568 loss) | |
| I0510 17:11:23.735579 10926 solver.cpp:245] Train net output #141: loss3/loss17 = 0.465753 (* 0.0909091 = 0.0423412 loss) | |
| I0510 17:11:23.735594 10926 solver.cpp:245] Train net output #142: loss3/loss18 = 0.615304 (* 0.0909091 = 0.0559367 loss) | |
| I0510 17:11:23.735608 10926 solver.cpp:245] Train net output #143: loss3/loss19 = 0.0257501 (* 0.0909091 = 0.00234092 loss) | |
| I0510 17:11:23.735622 10926 solver.cpp:245] Train net output #144: loss3/loss20 = 0.0115768 (* 0.0909091 = 0.00105243 loss) | |
| I0510 17:11:23.735636 10926 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00918084 (* 0.0909091 = 0.000834622 loss) | |
| I0510 17:11:23.735651 10926 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00281999 (* 0.0909091 = 0.000256362 loss) | |
| I0510 17:11:23.735662 10926 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0510 17:11:23.735674 10926 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0510 17:11:23.735687 10926 solver.cpp:245] Train net output #149: total_confidence = 1.55514e-05 | |
| I0510 17:11:23.735712 10926 solver.cpp:245] Train net output #150: total_confidence_not_rec = 3.63809e-05 | |
| I0510 17:11:23.735728 10926 sgd_solver.cpp:106] Iteration 20500, lr = 0.001 | |
| I0510 17:12:36.696514 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 30.0312 > 30) by scale factor 0.998962 | |
| I0510 17:13:36.695987 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 34.014 > 30) by scale factor 0.881991 | |
| I0510 17:13:50.995594 10926 solver.cpp:229] Iteration 21000, loss = 9.7342 | |
| I0510 17:13:50.995668 10926 solver.cpp:245] Train net output #0: loss1/accuracy = 0.136364 | |
| I0510 17:13:50.995703 10926 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.125 | |
| I0510 17:13:50.995730 10926 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0510 17:13:50.995754 10926 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.375 | |
| I0510 17:13:50.995776 10926 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0 | |
| I0510 17:13:50.995798 10926 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.125 | |
| I0510 17:13:50.995825 10926 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.625 | |
| I0510 17:13:50.995848 10926 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.75 | |
| I0510 17:13:50.995872 10926 solver.cpp:245] Train net output #8: loss1/accuracy08 = 1 | |
| I0510 17:13:50.995894 10926 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0510 17:13:50.995916 10926 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0510 17:13:50.995939 10926 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0510 17:13:50.995960 10926 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0510 17:13:50.995982 10926 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0510 17:13:50.996006 10926 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0510 17:13:50.996027 10926 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0510 17:13:50.996048 10926 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0510 17:13:50.996069 10926 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0510 17:13:50.996091 10926 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0510 17:13:50.996112 10926 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0510 17:13:50.996134 10926 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0510 17:13:50.996155 10926 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0510 17:13:50.996177 10926 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0510 17:13:50.996202 10926 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.767045 | |
| I0510 17:13:50.996227 10926 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.272727 | |
| I0510 17:13:50.996259 10926 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.04959 (* 0.3 = 0.914878 loss) | |
| I0510 17:13:50.996296 10926 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 0.894548 (* 0.3 = 0.268364 loss) | |
| I0510 17:13:50.996325 10926 solver.cpp:245] Train net output #27: loss1/loss01 = 3.29443 (* 0.0272727 = 0.0898481 loss) | |
| I0510 17:13:50.996352 10926 solver.cpp:245] Train net output #28: loss1/loss02 = 3.15345 (* 0.0272727 = 0.0860031 loss) | |
| I0510 17:13:50.996383 10926 solver.cpp:245] Train net output #29: loss1/loss03 = 2.93627 (* 0.0272727 = 0.08008 loss) | |
| I0510 17:13:50.996410 10926 solver.cpp:245] Train net output #30: loss1/loss04 = 3.41565 (* 0.0272727 = 0.0931541 loss) | |
| I0510 17:13:50.996436 10926 solver.cpp:245] Train net output #31: loss1/loss05 = 3.43038 (* 0.0272727 = 0.0935557 loss) | |
| I0510 17:13:50.996462 10926 solver.cpp:245] Train net output #32: loss1/loss06 = 1.76187 (* 0.0272727 = 0.048051 loss) | |
| I0510 17:13:50.996490 10926 solver.cpp:245] Train net output #33: loss1/loss07 = 1.11839 (* 0.0272727 = 0.0305016 loss) | |
| I0510 17:13:50.996516 10926 solver.cpp:245] Train net output #34: loss1/loss08 = 0.125741 (* 0.0272727 = 0.00342929 loss) | |
| I0510 17:13:50.996543 10926 solver.cpp:245] Train net output #35: loss1/loss09 = 0.0231143 (* 0.0272727 = 0.000630389 loss) | |
| I0510 17:13:50.996570 10926 solver.cpp:245] Train net output #36: loss1/loss10 = 0.0166209 (* 0.0272727 = 0.000453297 loss) | |
| I0510 17:13:50.996598 10926 solver.cpp:245] Train net output #37: loss1/loss11 = 0.0128956 (* 0.0272727 = 0.000351699 loss) | |
| I0510 17:13:50.996624 10926 solver.cpp:245] Train net output #38: loss1/loss12 = 0.00614798 (* 0.0272727 = 0.000167672 loss) | |
| I0510 17:13:50.996687 10926 solver.cpp:245] Train net output #39: loss1/loss13 = 0.00741483 (* 0.0272727 = 0.000202223 loss) | |
| I0510 17:13:50.996716 10926 solver.cpp:245] Train net output #40: loss1/loss14 = 0.00777176 (* 0.0272727 = 0.000211957 loss) | |
| I0510 17:13:50.996744 10926 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0101874 (* 0.0272727 = 0.000277837 loss) | |
| I0510 17:13:50.996775 10926 solver.cpp:245] Train net output #42: loss1/loss16 = 0.00899565 (* 0.0272727 = 0.000245336 loss) | |
| I0510 17:13:50.996803 10926 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0097867 (* 0.0272727 = 0.00026691 loss) | |
| I0510 17:13:50.996829 10926 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0106504 (* 0.0272727 = 0.000290465 loss) | |
| I0510 17:13:50.996855 10926 solver.cpp:245] Train net output #45: loss1/loss19 = 0.00668382 (* 0.0272727 = 0.000182286 loss) | |
| I0510 17:13:50.996882 10926 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00757356 (* 0.0272727 = 0.000206552 loss) | |
| I0510 17:13:50.996909 10926 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00791015 (* 0.0272727 = 0.000215731 loss) | |
| I0510 17:13:50.996937 10926 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0122801 (* 0.0272727 = 0.000334911 loss) | |
| I0510 17:13:50.996959 10926 solver.cpp:245] Train net output #49: loss2/accuracy = 0.113636 | |
| I0510 17:13:50.996983 10926 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.125 | |
| I0510 17:13:50.997004 10926 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0.25 | |
| I0510 17:13:50.997027 10926 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0.125 | |
| I0510 17:13:50.997050 10926 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.25 | |
| I0510 17:13:50.997071 10926 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.125 | |
| I0510 17:13:50.997092 10926 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.625 | |
| I0510 17:13:50.997117 10926 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.75 | |
| I0510 17:13:50.997155 10926 solver.cpp:245] Train net output #57: loss2/accuracy08 = 1 | |
| I0510 17:13:50.997179 10926 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0510 17:13:50.997202 10926 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0510 17:13:50.997223 10926 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0510 17:13:50.997246 10926 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0510 17:13:50.997267 10926 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0510 17:13:50.997288 10926 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0510 17:13:50.997310 10926 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0510 17:13:50.997333 10926 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0510 17:13:50.997354 10926 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0510 17:13:50.997375 10926 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0510 17:13:50.997397 10926 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0510 17:13:50.997419 10926 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0510 17:13:50.997440 10926 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0510 17:13:50.997457 10926 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0510 17:13:50.997481 10926 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.767045 | |
| I0510 17:13:50.997514 10926 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.272727 | |
| I0510 17:13:50.997542 10926 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 2.96681 (* 0.3 = 0.890043 loss) | |
| I0510 17:13:50.997570 10926 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 0.869883 (* 0.3 = 0.260965 loss) | |
| I0510 17:13:50.997597 10926 solver.cpp:245] Train net output #76: loss2/loss01 = 3.43611 (* 0.0272727 = 0.0937121 loss) | |
| I0510 17:13:50.997624 10926 solver.cpp:245] Train net output #77: loss2/loss02 = 2.85118 (* 0.0272727 = 0.0777594 loss) | |
| I0510 17:13:50.997670 10926 solver.cpp:245] Train net output #78: loss2/loss03 = 2.79284 (* 0.0272727 = 0.0761683 loss) | |
| I0510 17:13:50.997699 10926 solver.cpp:245] Train net output #79: loss2/loss04 = 3.38602 (* 0.0272727 = 0.0923459 loss) | |
| I0510 17:13:50.997725 10926 solver.cpp:245] Train net output #80: loss2/loss05 = 2.97358 (* 0.0272727 = 0.0810976 loss) | |
| I0510 17:13:50.997750 10926 solver.cpp:245] Train net output #81: loss2/loss06 = 1.66368 (* 0.0272727 = 0.0453732 loss) | |
| I0510 17:13:50.997776 10926 solver.cpp:245] Train net output #82: loss2/loss07 = 1.32338 (* 0.0272727 = 0.0360923 loss) | |
| I0510 17:13:50.997808 10926 solver.cpp:245] Train net output #83: loss2/loss08 = 0.079129 (* 0.0272727 = 0.00215806 loss) | |
| I0510 17:13:50.997834 10926 solver.cpp:245] Train net output #84: loss2/loss09 = 0.0224945 (* 0.0272727 = 0.000613485 loss) | |
| I0510 17:13:50.997860 10926 solver.cpp:245] Train net output #85: loss2/loss10 = 0.0128755 (* 0.0272727 = 0.000351149 loss) | |
| I0510 17:13:50.997887 10926 solver.cpp:245] Train net output #86: loss2/loss11 = 0.0118091 (* 0.0272727 = 0.000322067 loss) | |
| I0510 17:13:50.997913 10926 solver.cpp:245] Train net output #87: loss2/loss12 = 0.00875012 (* 0.0272727 = 0.00023864 loss) | |
| I0510 17:13:50.997941 10926 solver.cpp:245] Train net output #88: loss2/loss13 = 0.011426 (* 0.0272727 = 0.000311618 loss) | |
| I0510 17:13:50.997967 10926 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0192107 (* 0.0272727 = 0.000523927 loss) | |
| I0510 17:13:50.997994 10926 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0106079 (* 0.0272727 = 0.000289305 loss) | |
| I0510 17:13:50.998023 10926 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0107596 (* 0.0272727 = 0.000293444 loss) | |
| I0510 17:13:50.998049 10926 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0114471 (* 0.0272727 = 0.000312192 loss) | |
| I0510 17:13:50.998075 10926 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0137321 (* 0.0272727 = 0.000374513 loss) | |
| I0510 17:13:50.998100 10926 solver.cpp:245] Train net output #94: loss2/loss19 = 0.0187127 (* 0.0272727 = 0.000510346 loss) | |
| I0510 17:13:50.998126 10926 solver.cpp:245] Train net output #95: loss2/loss20 = 0.0223192 (* 0.0272727 = 0.000608705 loss) | |
| I0510 17:13:50.998152 10926 solver.cpp:245] Train net output #96: loss2/loss21 = 0.0139241 (* 0.0272727 = 0.000379749 loss) | |
| I0510 17:13:50.998178 10926 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0106054 (* 0.0272727 = 0.000289239 loss) | |
| I0510 17:13:50.998199 10926 solver.cpp:245] Train net output #98: loss3/accuracy = 0.113636 | |
| I0510 17:13:50.998222 10926 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.125 | |
| I0510 17:13:50.998244 10926 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.125 | |
| I0510 17:13:50.998265 10926 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0.25 | |
| I0510 17:13:50.998286 10926 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0510 17:13:50.998308 10926 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.375 | |
| I0510 17:13:50.998329 10926 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.625 | |
| I0510 17:13:50.998350 10926 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.75 | |
| I0510 17:13:50.998371 10926 solver.cpp:245] Train net output #106: loss3/accuracy08 = 1 | |
| I0510 17:13:50.998392 10926 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0510 17:13:50.998414 10926 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0510 17:13:50.998435 10926 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0510 17:13:50.998456 10926 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0510 17:13:50.998482 10926 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0510 17:13:50.998504 10926 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0510 17:13:50.998525 10926 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0510 17:13:50.998563 10926 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0510 17:13:50.998586 10926 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0510 17:13:50.998606 10926 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0510 17:13:50.998630 10926 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0510 17:13:50.998651 10926 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0510 17:13:50.998672 10926 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0510 17:13:50.998692 10926 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0510 17:13:50.998713 10926 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.772727 | |
| I0510 17:13:50.998736 10926 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.318182 | |
| I0510 17:13:50.998761 10926 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 2.91508 (* 1 = 2.91508 loss) | |
| I0510 17:13:50.998787 10926 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 0.794968 (* 1 = 0.794968 loss) | |
| I0510 17:13:50.998816 10926 solver.cpp:245] Train net output #125: loss3/loss01 = 3.08425 (* 0.0909091 = 0.280387 loss) | |
| I0510 17:13:50.998842 10926 solver.cpp:245] Train net output #126: loss3/loss02 = 3.00288 (* 0.0909091 = 0.272989 loss) | |
| I0510 17:13:50.998872 10926 solver.cpp:245] Train net output #127: loss3/loss03 = 2.70068 (* 0.0909091 = 0.245516 loss) | |
| I0510 17:13:50.998898 10926 solver.cpp:245] Train net output #128: loss3/loss04 = 3.12286 (* 0.0909091 = 0.283896 loss) | |
| I0510 17:13:50.998922 10926 solver.cpp:245] Train net output #129: loss3/loss05 = 2.80191 (* 0.0909091 = 0.254719 loss) | |
| I0510 17:13:50.998949 10926 solver.cpp:245] Train net output #130: loss3/loss06 = 1.57929 (* 0.0909091 = 0.143572 loss) | |
| I0510 17:13:50.998975 10926 solver.cpp:245] Train net output #131: loss3/loss07 = 1.01431 (* 0.0909091 = 0.0922103 loss) | |
| I0510 17:13:50.999001 10926 solver.cpp:245] Train net output #132: loss3/loss08 = 0.0607765 (* 0.0909091 = 0.00552514 loss) | |
| I0510 17:13:50.999027 10926 solver.cpp:245] Train net output #133: loss3/loss09 = 0.0117735 (* 0.0909091 = 0.00107032 loss) | |
| I0510 17:13:50.999054 10926 solver.cpp:245] Train net output #134: loss3/loss10 = 0.00433071 (* 0.0909091 = 0.000393701 loss) | |
| I0510 17:13:50.999080 10926 solver.cpp:245] Train net output #135: loss3/loss11 = 0.00231639 (* 0.0909091 = 0.000210581 loss) | |
| I0510 17:13:50.999105 10926 solver.cpp:245] Train net output #136: loss3/loss12 = 0.00249229 (* 0.0909091 = 0.000226572 loss) | |
| I0510 17:13:50.999131 10926 solver.cpp:245] Train net output #137: loss3/loss13 = 0.00147428 (* 0.0909091 = 0.000134025 loss) | |
| I0510 17:13:50.999158 10926 solver.cpp:245] Train net output #138: loss3/loss14 = 0.0013218 (* 0.0909091 = 0.000120164 loss) | |
| I0510 17:13:50.999184 10926 solver.cpp:245] Train net output #139: loss3/loss15 = 0.00136988 (* 0.0909091 = 0.000124534 loss) | |
| I0510 17:13:50.999209 10926 solver.cpp:245] Train net output #140: loss3/loss16 = 0.00131769 (* 0.0909091 = 0.00011979 loss) | |
| I0510 17:13:50.999236 10926 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00299258 (* 0.0909091 = 0.000272053 loss) | |
| I0510 17:13:50.999261 10926 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00269358 (* 0.0909091 = 0.000244871 loss) | |
| I0510 17:13:50.999289 10926 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00266556 (* 0.0909091 = 0.000242324 loss) | |
| I0510 17:13:50.999315 10926 solver.cpp:245] Train net output #144: loss3/loss20 = 0.0027882 (* 0.0909091 = 0.000253473 loss) | |
| I0510 17:13:50.999341 10926 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00189677 (* 0.0909091 = 0.000172434 loss) | |
| I0510 17:13:50.999366 10926 solver.cpp:245] Train net output #146: loss3/loss22 = 0.0019658 (* 0.0909091 = 0.000178709 loss) | |
| I0510 17:13:50.999388 10926 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0510 17:13:50.999409 10926 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0510 17:13:50.999446 10926 solver.cpp:245] Train net output #149: total_confidence = 2.16333e-06 | |
| I0510 17:13:50.999470 10926 solver.cpp:245] Train net output #150: total_confidence_not_rec = 7.86322e-06 | |
| I0510 17:13:50.999495 10926 sgd_solver.cpp:106] Iteration 21000, lr = 0.001 | |
| I0510 17:14:18.479445 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 39.5007 > 30) by scale factor 0.75948 | |
| I0510 17:15:25.773538 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 33.3333 > 30) by scale factor 0.9 | |
| I0510 17:15:45.192101 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 39.9205 > 30) by scale factor 0.751494 | |
| I0510 17:16:07.887511 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 43.5447 > 30) by scale factor 0.688947 | |
| I0510 17:16:18.080312 10926 solver.cpp:229] Iteration 21500, loss = 9.75013 | |
| I0510 17:16:18.080364 10926 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0535714 | |
| I0510 17:16:18.080384 10926 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.25 | |
| I0510 17:16:18.080397 10926 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0510 17:16:18.080410 10926 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0510 17:16:18.080422 10926 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0 | |
| I0510 17:16:18.080435 10926 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.375 | |
| I0510 17:16:18.080448 10926 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.375 | |
| I0510 17:16:18.080461 10926 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.625 | |
| I0510 17:16:18.080473 10926 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.75 | |
| I0510 17:16:18.080487 10926 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0510 17:16:18.080499 10926 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.875 | |
| I0510 17:16:18.080513 10926 solver.cpp:245] Train net output #11: loss1/accuracy11 = 0.875 | |
| I0510 17:16:18.080524 10926 solver.cpp:245] Train net output #12: loss1/accuracy12 = 0.875 | |
| I0510 17:16:18.080538 10926 solver.cpp:245] Train net output #13: loss1/accuracy13 = 0.875 | |
| I0510 17:16:18.080549 10926 solver.cpp:245] Train net output #14: loss1/accuracy14 = 0.875 | |
| I0510 17:16:18.080561 10926 solver.cpp:245] Train net output #15: loss1/accuracy15 = 0.875 | |
| I0510 17:16:18.080574 10926 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0510 17:16:18.080585 10926 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0510 17:16:18.080597 10926 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0510 17:16:18.080608 10926 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0510 17:16:18.080621 10926 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0510 17:16:18.080636 10926 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0510 17:16:18.080647 10926 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0510 17:16:18.080659 10926 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.693182 | |
| I0510 17:16:18.080672 10926 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.25 | |
| I0510 17:16:18.080687 10926 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.42771 (* 0.3 = 1.02831 loss) | |
| I0510 17:16:18.080703 10926 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.18755 (* 0.3 = 0.356266 loss) | |
| I0510 17:16:18.080718 10926 solver.cpp:245] Train net output #27: loss1/loss01 = 3.31922 (* 0.0272727 = 0.0905243 loss) | |
| I0510 17:16:18.080731 10926 solver.cpp:245] Train net output #28: loss1/loss02 = 3.46759 (* 0.0272727 = 0.0945706 loss) | |
| I0510 17:16:18.080746 10926 solver.cpp:245] Train net output #29: loss1/loss03 = 3.47517 (* 0.0272727 = 0.0947775 loss) | |
| I0510 17:16:18.080760 10926 solver.cpp:245] Train net output #30: loss1/loss04 = 3.21463 (* 0.0272727 = 0.0876718 loss) | |
| I0510 17:16:18.080775 10926 solver.cpp:245] Train net output #31: loss1/loss05 = 2.80546 (* 0.0272727 = 0.0765126 loss) | |
| I0510 17:16:18.080788 10926 solver.cpp:245] Train net output #32: loss1/loss06 = 2.8242 (* 0.0272727 = 0.0770235 loss) | |
| I0510 17:16:18.080806 10926 solver.cpp:245] Train net output #33: loss1/loss07 = 1.97674 (* 0.0272727 = 0.0539111 loss) | |
| I0510 17:16:18.080821 10926 solver.cpp:245] Train net output #34: loss1/loss08 = 0.976525 (* 0.0272727 = 0.0266325 loss) | |
| I0510 17:16:18.080837 10926 solver.cpp:245] Train net output #35: loss1/loss09 = 0.28322 (* 0.0272727 = 0.00772418 loss) | |
| I0510 17:16:18.080850 10926 solver.cpp:245] Train net output #36: loss1/loss10 = 0.465182 (* 0.0272727 = 0.0126868 loss) | |
| I0510 17:16:18.080864 10926 solver.cpp:245] Train net output #37: loss1/loss11 = 0.417735 (* 0.0272727 = 0.0113928 loss) | |
| I0510 17:16:18.080879 10926 solver.cpp:245] Train net output #38: loss1/loss12 = 0.496411 (* 0.0272727 = 0.0135385 loss) | |
| I0510 17:16:18.080924 10926 solver.cpp:245] Train net output #39: loss1/loss13 = 0.396055 (* 0.0272727 = 0.0108015 loss) | |
| I0510 17:16:18.080941 10926 solver.cpp:245] Train net output #40: loss1/loss14 = 0.553645 (* 0.0272727 = 0.0150994 loss) | |
| I0510 17:16:18.080955 10926 solver.cpp:245] Train net output #41: loss1/loss15 = 0.482632 (* 0.0272727 = 0.0131627 loss) | |
| I0510 17:16:18.080971 10926 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0258861 (* 0.0272727 = 0.000705983 loss) | |
| I0510 17:16:18.080984 10926 solver.cpp:245] Train net output #43: loss1/loss17 = 0.012747 (* 0.0272727 = 0.000347647 loss) | |
| I0510 17:16:18.080998 10926 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0141359 (* 0.0272727 = 0.000385525 loss) | |
| I0510 17:16:18.081013 10926 solver.cpp:245] Train net output #45: loss1/loss19 = 0.00352509 (* 0.0272727 = 9.61388e-05 loss) | |
| I0510 17:16:18.081027 10926 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0129432 (* 0.0272727 = 0.000352997 loss) | |
| I0510 17:16:18.081043 10926 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00361267 (* 0.0272727 = 9.85274e-05 loss) | |
| I0510 17:16:18.081056 10926 solver.cpp:245] Train net output #48: loss1/loss22 = 0.00390434 (* 0.0272727 = 0.000106482 loss) | |
| I0510 17:16:18.081068 10926 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0535714 | |
| I0510 17:16:18.081080 10926 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0510 17:16:18.081092 10926 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0.125 | |
| I0510 17:16:18.081105 10926 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0510 17:16:18.081128 10926 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.125 | |
| I0510 17:16:18.081145 10926 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.5 | |
| I0510 17:16:18.081156 10926 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.375 | |
| I0510 17:16:18.081169 10926 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.625 | |
| I0510 17:16:18.081182 10926 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.75 | |
| I0510 17:16:18.081193 10926 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.875 | |
| I0510 17:16:18.081204 10926 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.875 | |
| I0510 17:16:18.081218 10926 solver.cpp:245] Train net output #60: loss2/accuracy11 = 0.875 | |
| I0510 17:16:18.081229 10926 solver.cpp:245] Train net output #61: loss2/accuracy12 = 0.875 | |
| I0510 17:16:18.081240 10926 solver.cpp:245] Train net output #62: loss2/accuracy13 = 0.875 | |
| I0510 17:16:18.081253 10926 solver.cpp:245] Train net output #63: loss2/accuracy14 = 0.875 | |
| I0510 17:16:18.081264 10926 solver.cpp:245] Train net output #64: loss2/accuracy15 = 0.875 | |
| I0510 17:16:18.081276 10926 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0510 17:16:18.081287 10926 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0510 17:16:18.081298 10926 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0510 17:16:18.081310 10926 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0510 17:16:18.081321 10926 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0510 17:16:18.081332 10926 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0510 17:16:18.081344 10926 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0510 17:16:18.081356 10926 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.6875 | |
| I0510 17:16:18.081368 10926 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.285714 | |
| I0510 17:16:18.081382 10926 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.33008 (* 0.3 = 0.999024 loss) | |
| I0510 17:16:18.081396 10926 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.1634 (* 0.3 = 0.349019 loss) | |
| I0510 17:16:18.081410 10926 solver.cpp:245] Train net output #76: loss2/loss01 = 3.46394 (* 0.0272727 = 0.0944711 loss) | |
| I0510 17:16:18.081425 10926 solver.cpp:245] Train net output #77: loss2/loss02 = 3.49622 (* 0.0272727 = 0.0953513 loss) | |
| I0510 17:16:18.081451 10926 solver.cpp:245] Train net output #78: loss2/loss03 = 3.59219 (* 0.0272727 = 0.0979687 loss) | |
| I0510 17:16:18.081466 10926 solver.cpp:245] Train net output #79: loss2/loss04 = 3.5226 (* 0.0272727 = 0.0960708 loss) | |
| I0510 17:16:18.081481 10926 solver.cpp:245] Train net output #80: loss2/loss05 = 2.62373 (* 0.0272727 = 0.0715563 loss) | |
| I0510 17:16:18.081496 10926 solver.cpp:245] Train net output #81: loss2/loss06 = 2.8785 (* 0.0272727 = 0.0785046 loss) | |
| I0510 17:16:18.081509 10926 solver.cpp:245] Train net output #82: loss2/loss07 = 1.73105 (* 0.0272727 = 0.0472103 loss) | |
| I0510 17:16:18.081523 10926 solver.cpp:245] Train net output #83: loss2/loss08 = 1.11178 (* 0.0272727 = 0.0303214 loss) | |
| I0510 17:16:18.081537 10926 solver.cpp:245] Train net output #84: loss2/loss09 = 0.433861 (* 0.0272727 = 0.0118326 loss) | |
| I0510 17:16:18.081552 10926 solver.cpp:245] Train net output #85: loss2/loss10 = 0.463459 (* 0.0272727 = 0.0126398 loss) | |
| I0510 17:16:18.081567 10926 solver.cpp:245] Train net output #86: loss2/loss11 = 0.424945 (* 0.0272727 = 0.0115894 loss) | |
| I0510 17:16:18.081581 10926 solver.cpp:245] Train net output #87: loss2/loss12 = 0.32788 (* 0.0272727 = 0.00894219 loss) | |
| I0510 17:16:18.081593 10926 solver.cpp:245] Train net output #88: loss2/loss13 = 0.483537 (* 0.0272727 = 0.0131874 loss) | |
| I0510 17:16:18.081601 10926 solver.cpp:245] Train net output #89: loss2/loss14 = 0.57819 (* 0.0272727 = 0.0157688 loss) | |
| I0510 17:16:18.081616 10926 solver.cpp:245] Train net output #90: loss2/loss15 = 0.547256 (* 0.0272727 = 0.0149252 loss) | |
| I0510 17:16:18.081630 10926 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0171735 (* 0.0272727 = 0.000468367 loss) | |
| I0510 17:16:18.081645 10926 solver.cpp:245] Train net output #92: loss2/loss17 = 0.00466409 (* 0.0272727 = 0.000127202 loss) | |
| I0510 17:16:18.081660 10926 solver.cpp:245] Train net output #93: loss2/loss18 = 0.00753478 (* 0.0272727 = 0.000205494 loss) | |
| I0510 17:16:18.081676 10926 solver.cpp:245] Train net output #94: loss2/loss19 = 0.00442191 (* 0.0272727 = 0.000120597 loss) | |
| I0510 17:16:18.081691 10926 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00515928 (* 0.0272727 = 0.000140708 loss) | |
| I0510 17:16:18.081706 10926 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00206883 (* 0.0272727 = 5.64225e-05 loss) | |
| I0510 17:16:18.081719 10926 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0020882 (* 0.0272727 = 5.69509e-05 loss) | |
| I0510 17:16:18.081732 10926 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0714286 | |
| I0510 17:16:18.081744 10926 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.125 | |
| I0510 17:16:18.081756 10926 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.125 | |
| I0510 17:16:18.081768 10926 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0510 17:16:18.081780 10926 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.25 | |
| I0510 17:16:18.081792 10926 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.625 | |
| I0510 17:16:18.081804 10926 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.375 | |
| I0510 17:16:18.081815 10926 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.625 | |
| I0510 17:16:18.081827 10926 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.75 | |
| I0510 17:16:18.081838 10926 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.875 | |
| I0510 17:16:18.081850 10926 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.875 | |
| I0510 17:16:18.081864 10926 solver.cpp:245] Train net output #109: loss3/accuracy11 = 0.875 | |
| I0510 17:16:18.081877 10926 solver.cpp:245] Train net output #110: loss3/accuracy12 = 0.875 | |
| I0510 17:16:18.081888 10926 solver.cpp:245] Train net output #111: loss3/accuracy13 = 0.875 | |
| I0510 17:16:18.081900 10926 solver.cpp:245] Train net output #112: loss3/accuracy14 = 0.875 | |
| I0510 17:16:18.081912 10926 solver.cpp:245] Train net output #113: loss3/accuracy15 = 0.875 | |
| I0510 17:16:18.081933 10926 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0510 17:16:18.081946 10926 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0510 17:16:18.081959 10926 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0510 17:16:18.081970 10926 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0510 17:16:18.081981 10926 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0510 17:16:18.081993 10926 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0510 17:16:18.082005 10926 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0510 17:16:18.082016 10926 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.698864 | |
| I0510 17:16:18.082028 10926 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.25 | |
| I0510 17:16:18.082041 10926 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.32196 (* 1 = 3.32196 loss) | |
| I0510 17:16:18.082056 10926 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.13654 (* 1 = 1.13654 loss) | |
| I0510 17:16:18.082069 10926 solver.cpp:245] Train net output #125: loss3/loss01 = 3.16743 (* 0.0909091 = 0.287948 loss) | |
| I0510 17:16:18.082083 10926 solver.cpp:245] Train net output #126: loss3/loss02 = 3.38966 (* 0.0909091 = 0.308151 loss) | |
| I0510 17:16:18.082098 10926 solver.cpp:245] Train net output #127: loss3/loss03 = 3.52703 (* 0.0909091 = 0.320639 loss) | |
| I0510 17:16:18.082111 10926 solver.cpp:245] Train net output #128: loss3/loss04 = 3.44953 (* 0.0909091 = 0.313594 loss) | |
| I0510 17:16:18.082125 10926 solver.cpp:245] Train net output #129: loss3/loss05 = 2.41202 (* 0.0909091 = 0.219275 loss) | |
| I0510 17:16:18.082139 10926 solver.cpp:245] Train net output #130: loss3/loss06 = 2.58798 (* 0.0909091 = 0.235271 loss) | |
| I0510 17:16:18.082154 10926 solver.cpp:245] Train net output #131: loss3/loss07 = 1.81788 (* 0.0909091 = 0.165262 loss) | |
| I0510 17:16:18.082167 10926 solver.cpp:245] Train net output #132: loss3/loss08 = 0.817999 (* 0.0909091 = 0.0743635 loss) | |
| I0510 17:16:18.082181 10926 solver.cpp:245] Train net output #133: loss3/loss09 = 0.381705 (* 0.0909091 = 0.0347005 loss) | |
| I0510 17:16:18.082195 10926 solver.cpp:245] Train net output #134: loss3/loss10 = 0.357358 (* 0.0909091 = 0.0324871 loss) | |
| I0510 17:16:18.082209 10926 solver.cpp:245] Train net output #135: loss3/loss11 = 0.488448 (* 0.0909091 = 0.0444044 loss) | |
| I0510 17:16:18.082223 10926 solver.cpp:245] Train net output #136: loss3/loss12 = 0.406724 (* 0.0909091 = 0.0369749 loss) | |
| I0510 17:16:18.082238 10926 solver.cpp:245] Train net output #137: loss3/loss13 = 0.476269 (* 0.0909091 = 0.0432972 loss) | |
| I0510 17:16:18.082252 10926 solver.cpp:245] Train net output #138: loss3/loss14 = 0.452349 (* 0.0909091 = 0.0411226 loss) | |
| I0510 17:16:18.082265 10926 solver.cpp:245] Train net output #139: loss3/loss15 = 0.580102 (* 0.0909091 = 0.0527366 loss) | |
| I0510 17:16:18.082279 10926 solver.cpp:245] Train net output #140: loss3/loss16 = 0.00653462 (* 0.0909091 = 0.000594057 loss) | |
| I0510 17:16:18.082294 10926 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00736518 (* 0.0909091 = 0.000669562 loss) | |
| I0510 17:16:18.082309 10926 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00359128 (* 0.0909091 = 0.00032648 loss) | |
| I0510 17:16:18.082322 10926 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00338928 (* 0.0909091 = 0.000308116 loss) | |
| I0510 17:16:18.082337 10926 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00147837 (* 0.0909091 = 0.000134397 loss) | |
| I0510 17:16:18.082351 10926 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00166274 (* 0.0909091 = 0.000151158 loss) | |
| I0510 17:16:18.082365 10926 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00080216 (* 0.0909091 = 7.29237e-05 loss) | |
| I0510 17:16:18.082377 10926 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0510 17:16:18.082389 10926 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0510 17:16:18.082411 10926 solver.cpp:245] Train net output #149: total_confidence = 3.48978e-06 | |
| I0510 17:16:18.082423 10926 solver.cpp:245] Train net output #150: total_confidence_not_rec = 8.31119e-05 | |
| I0510 17:16:18.082437 10926 sgd_solver.cpp:106] Iteration 21500, lr = 0.001 | |
| I0510 17:18:43.541759 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 51.7416 > 30) by scale factor 0.579804 | |
| I0510 17:18:45.510576 10926 solver.cpp:229] Iteration 22000, loss = 9.73787 | |
| I0510 17:18:45.510635 10926 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0892857 | |
| I0510 17:18:45.510658 10926 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0510 17:18:45.510673 10926 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.125 | |
| I0510 17:18:45.510685 10926 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.25 | |
| I0510 17:18:45.510699 10926 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0 | |
| I0510 17:18:45.510711 10926 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.125 | |
| I0510 17:18:45.510725 10926 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.125 | |
| I0510 17:18:45.510738 10926 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.375 | |
| I0510 17:18:45.510753 10926 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.625 | |
| I0510 17:18:45.510767 10926 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.875 | |
| I0510 17:18:45.510781 10926 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.875 | |
| I0510 17:18:45.510792 10926 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0510 17:18:45.510805 10926 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0510 17:18:45.510818 10926 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0510 17:18:45.510830 10926 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0510 17:18:45.510843 10926 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0510 17:18:45.510854 10926 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0510 17:18:45.510866 10926 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0510 17:18:45.510879 10926 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0510 17:18:45.510890 10926 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0510 17:18:45.510902 10926 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0510 17:18:45.510913 10926 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0510 17:18:45.510926 10926 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0510 17:18:45.510938 10926 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.710227 | |
| I0510 17:18:45.510951 10926 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.142857 | |
| I0510 17:18:45.510967 10926 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.33962 (* 0.3 = 1.00189 loss) | |
| I0510 17:18:45.510982 10926 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.1064 (* 0.3 = 0.331921 loss) | |
| I0510 17:18:45.510995 10926 solver.cpp:245] Train net output #27: loss1/loss01 = 3.29203 (* 0.0272727 = 0.0897826 loss) | |
| I0510 17:18:45.511010 10926 solver.cpp:245] Train net output #28: loss1/loss02 = 3.23703 (* 0.0272727 = 0.0882825 loss) | |
| I0510 17:18:45.511024 10926 solver.cpp:245] Train net output #29: loss1/loss03 = 3.21108 (* 0.0272727 = 0.087575 loss) | |
| I0510 17:18:45.511039 10926 solver.cpp:245] Train net output #30: loss1/loss04 = 3.35029 (* 0.0272727 = 0.0913714 loss) | |
| I0510 17:18:45.511052 10926 solver.cpp:245] Train net output #31: loss1/loss05 = 3.43616 (* 0.0272727 = 0.0937134 loss) | |
| I0510 17:18:45.511067 10926 solver.cpp:245] Train net output #32: loss1/loss06 = 3.5885 (* 0.0272727 = 0.0978682 loss) | |
| I0510 17:18:45.511081 10926 solver.cpp:245] Train net output #33: loss1/loss07 = 2.7866 (* 0.0272727 = 0.0759983 loss) | |
| I0510 17:18:45.511096 10926 solver.cpp:245] Train net output #34: loss1/loss08 = 1.23283 (* 0.0272727 = 0.0336227 loss) | |
| I0510 17:18:45.511111 10926 solver.cpp:245] Train net output #35: loss1/loss09 = 1.13607 (* 0.0272727 = 0.0309836 loss) | |
| I0510 17:18:45.511126 10926 solver.cpp:245] Train net output #36: loss1/loss10 = 0.6169 (* 0.0272727 = 0.0168245 loss) | |
| I0510 17:18:45.511139 10926 solver.cpp:245] Train net output #37: loss1/loss11 = 0.0394189 (* 0.0272727 = 0.00107506 loss) | |
| I0510 17:18:45.511154 10926 solver.cpp:245] Train net output #38: loss1/loss12 = 0.0440906 (* 0.0272727 = 0.00120247 loss) | |
| I0510 17:18:45.511209 10926 solver.cpp:245] Train net output #39: loss1/loss13 = 0.0238588 (* 0.0272727 = 0.000650694 loss) | |
| I0510 17:18:45.511226 10926 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0238083 (* 0.0272727 = 0.000649319 loss) | |
| I0510 17:18:45.511240 10926 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0214443 (* 0.0272727 = 0.000584846 loss) | |
| I0510 17:18:45.511255 10926 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0174493 (* 0.0272727 = 0.000475889 loss) | |
| I0510 17:18:45.511270 10926 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0120223 (* 0.0272727 = 0.00032788 loss) | |
| I0510 17:18:45.511283 10926 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0085321 (* 0.0272727 = 0.000232694 loss) | |
| I0510 17:18:45.511298 10926 solver.cpp:245] Train net output #45: loss1/loss19 = 0.00566834 (* 0.0272727 = 0.000154591 loss) | |
| I0510 17:18:45.511312 10926 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00716218 (* 0.0272727 = 0.000195332 loss) | |
| I0510 17:18:45.511327 10926 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0080759 (* 0.0272727 = 0.000220252 loss) | |
| I0510 17:18:45.511342 10926 solver.cpp:245] Train net output #48: loss1/loss22 = 0.00303278 (* 0.0272727 = 8.27123e-05 loss) | |
| I0510 17:18:45.511353 10926 solver.cpp:245] Train net output #49: loss2/accuracy = 0 | |
| I0510 17:18:45.511365 10926 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0510 17:18:45.511378 10926 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0.125 | |
| I0510 17:18:45.511389 10926 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0510 17:18:45.511400 10926 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.125 | |
| I0510 17:18:45.511414 10926 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.125 | |
| I0510 17:18:45.511425 10926 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.125 | |
| I0510 17:18:45.511436 10926 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.375 | |
| I0510 17:18:45.511448 10926 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.625 | |
| I0510 17:18:45.511461 10926 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.875 | |
| I0510 17:18:45.511472 10926 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.875 | |
| I0510 17:18:45.511484 10926 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0510 17:18:45.511495 10926 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0510 17:18:45.511507 10926 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0510 17:18:45.511518 10926 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0510 17:18:45.511530 10926 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0510 17:18:45.511541 10926 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0510 17:18:45.511554 10926 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0510 17:18:45.511564 10926 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0510 17:18:45.511576 10926 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0510 17:18:45.511587 10926 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0510 17:18:45.511600 10926 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0510 17:18:45.511610 10926 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0510 17:18:45.511622 10926 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.664773 | |
| I0510 17:18:45.511633 10926 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.214286 | |
| I0510 17:18:45.511647 10926 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.25982 (* 0.3 = 0.977947 loss) | |
| I0510 17:18:45.511662 10926 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.18236 (* 0.3 = 0.354709 loss) | |
| I0510 17:18:45.511675 10926 solver.cpp:245] Train net output #76: loss2/loss01 = 3.23065 (* 0.0272727 = 0.0881087 loss) | |
| I0510 17:18:45.511690 10926 solver.cpp:245] Train net output #77: loss2/loss02 = 3.05926 (* 0.0272727 = 0.0834343 loss) | |
| I0510 17:18:45.511719 10926 solver.cpp:245] Train net output #78: loss2/loss03 = 3.43579 (* 0.0272727 = 0.0937033 loss) | |
| I0510 17:18:45.511734 10926 solver.cpp:245] Train net output #79: loss2/loss04 = 3.84598 (* 0.0272727 = 0.10489 loss) | |
| I0510 17:18:45.511749 10926 solver.cpp:245] Train net output #80: loss2/loss05 = 3.45074 (* 0.0272727 = 0.0941111 loss) | |
| I0510 17:18:45.511764 10926 solver.cpp:245] Train net output #81: loss2/loss06 = 3.37733 (* 0.0272727 = 0.0921089 loss) | |
| I0510 17:18:45.511777 10926 solver.cpp:245] Train net output #82: loss2/loss07 = 2.59663 (* 0.0272727 = 0.0708172 loss) | |
| I0510 17:18:45.511791 10926 solver.cpp:245] Train net output #83: loss2/loss08 = 1.20324 (* 0.0272727 = 0.0328157 loss) | |
| I0510 17:18:45.511807 10926 solver.cpp:245] Train net output #84: loss2/loss09 = 1.47627 (* 0.0272727 = 0.0402618 loss) | |
| I0510 17:18:45.511822 10926 solver.cpp:245] Train net output #85: loss2/loss10 = 0.826419 (* 0.0272727 = 0.0225387 loss) | |
| I0510 17:18:45.511837 10926 solver.cpp:245] Train net output #86: loss2/loss11 = 0.0942375 (* 0.0272727 = 0.00257011 loss) | |
| I0510 17:18:45.511852 10926 solver.cpp:245] Train net output #87: loss2/loss12 = 0.0940907 (* 0.0272727 = 0.00256611 loss) | |
| I0510 17:18:45.511865 10926 solver.cpp:245] Train net output #88: loss2/loss13 = 0.0617382 (* 0.0272727 = 0.00168377 loss) | |
| I0510 17:18:45.511879 10926 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0728301 (* 0.0272727 = 0.00198627 loss) | |
| I0510 17:18:45.511893 10926 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0739066 (* 0.0272727 = 0.00201564 loss) | |
| I0510 17:18:45.511907 10926 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0428581 (* 0.0272727 = 0.00116886 loss) | |
| I0510 17:18:45.511921 10926 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0215785 (* 0.0272727 = 0.000588504 loss) | |
| I0510 17:18:45.511935 10926 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0216817 (* 0.0272727 = 0.00059132 loss) | |
| I0510 17:18:45.511950 10926 solver.cpp:245] Train net output #94: loss2/loss19 = 0.0132151 (* 0.0272727 = 0.000360412 loss) | |
| I0510 17:18:45.511963 10926 solver.cpp:245] Train net output #95: loss2/loss20 = 0.02003 (* 0.0272727 = 0.000546273 loss) | |
| I0510 17:18:45.511977 10926 solver.cpp:245] Train net output #96: loss2/loss21 = 0.01031 (* 0.0272727 = 0.000281181 loss) | |
| I0510 17:18:45.511991 10926 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00616061 (* 0.0272727 = 0.000168017 loss) | |
| I0510 17:18:45.512004 10926 solver.cpp:245] Train net output #98: loss3/accuracy = 0.125 | |
| I0510 17:18:45.512017 10926 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0510 17:18:45.512028 10926 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0510 17:18:45.512040 10926 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0.25 | |
| I0510 17:18:45.512051 10926 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0 | |
| I0510 17:18:45.512063 10926 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.125 | |
| I0510 17:18:45.512075 10926 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.125 | |
| I0510 17:18:45.512086 10926 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.375 | |
| I0510 17:18:45.512099 10926 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.625 | |
| I0510 17:18:45.512109 10926 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.875 | |
| I0510 17:18:45.512121 10926 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.875 | |
| I0510 17:18:45.512133 10926 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0510 17:18:45.512145 10926 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0510 17:18:45.512156 10926 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0510 17:18:45.512168 10926 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0510 17:18:45.512181 10926 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0510 17:18:45.512198 10926 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0510 17:18:45.512212 10926 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0510 17:18:45.512225 10926 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0510 17:18:45.512236 10926 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0510 17:18:45.512248 10926 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0510 17:18:45.512259 10926 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0510 17:18:45.512270 10926 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0510 17:18:45.512281 10926 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.721591 | |
| I0510 17:18:45.512293 10926 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.196429 | |
| I0510 17:18:45.512307 10926 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.19006 (* 1 = 3.19006 loss) | |
| I0510 17:18:45.512321 10926 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.04742 (* 1 = 1.04742 loss) | |
| I0510 17:18:45.512336 10926 solver.cpp:245] Train net output #125: loss3/loss01 = 3.0119 (* 0.0909091 = 0.273809 loss) | |
| I0510 17:18:45.512348 10926 solver.cpp:245] Train net output #126: loss3/loss02 = 2.81201 (* 0.0909091 = 0.255637 loss) | |
| I0510 17:18:45.512362 10926 solver.cpp:245] Train net output #127: loss3/loss03 = 3.00438 (* 0.0909091 = 0.273126 loss) | |
| I0510 17:18:45.512377 10926 solver.cpp:245] Train net output #128: loss3/loss04 = 3.06671 (* 0.0909091 = 0.278792 loss) | |
| I0510 17:18:45.512390 10926 solver.cpp:245] Train net output #129: loss3/loss05 = 3.11897 (* 0.0909091 = 0.283543 loss) | |
| I0510 17:18:45.512403 10926 solver.cpp:245] Train net output #130: loss3/loss06 = 2.93987 (* 0.0909091 = 0.267261 loss) | |
| I0510 17:18:45.512418 10926 solver.cpp:245] Train net output #131: loss3/loss07 = 2.33525 (* 0.0909091 = 0.212295 loss) | |
| I0510 17:18:45.512431 10926 solver.cpp:245] Train net output #132: loss3/loss08 = 1.19354 (* 0.0909091 = 0.108503 loss) | |
| I0510 17:18:45.512444 10926 solver.cpp:245] Train net output #133: loss3/loss09 = 1.04834 (* 0.0909091 = 0.0953035 loss) | |
| I0510 17:18:45.512459 10926 solver.cpp:245] Train net output #134: loss3/loss10 = 0.816079 (* 0.0909091 = 0.074189 loss) | |
| I0510 17:18:45.512472 10926 solver.cpp:245] Train net output #135: loss3/loss11 = 0.034852 (* 0.0909091 = 0.00316836 loss) | |
| I0510 17:18:45.512487 10926 solver.cpp:245] Train net output #136: loss3/loss12 = 0.0190233 (* 0.0909091 = 0.00172939 loss) | |
| I0510 17:18:45.512501 10926 solver.cpp:245] Train net output #137: loss3/loss13 = 0.0136775 (* 0.0909091 = 0.00124341 loss) | |
| I0510 17:18:45.512514 10926 solver.cpp:245] Train net output #138: loss3/loss14 = 0.0123526 (* 0.0909091 = 0.00112296 loss) | |
| I0510 17:18:45.512528 10926 solver.cpp:245] Train net output #139: loss3/loss15 = 0.00806941 (* 0.0909091 = 0.000733583 loss) | |
| I0510 17:18:45.512542 10926 solver.cpp:245] Train net output #140: loss3/loss16 = 0.0065905 (* 0.0909091 = 0.000599137 loss) | |
| I0510 17:18:45.512557 10926 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00290518 (* 0.0909091 = 0.000264107 loss) | |
| I0510 17:18:45.512569 10926 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00262075 (* 0.0909091 = 0.00023825 loss) | |
| I0510 17:18:45.512583 10926 solver.cpp:245] Train net output #143: loss3/loss19 = 0.0017295 (* 0.0909091 = 0.000157227 loss) | |
| I0510 17:18:45.512598 10926 solver.cpp:245] Train net output #144: loss3/loss20 = 0.0018234 (* 0.0909091 = 0.000165763 loss) | |
| I0510 17:18:45.512611 10926 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00173799 (* 0.0909091 = 0.000157999 loss) | |
| I0510 17:18:45.512625 10926 solver.cpp:245] Train net output #146: loss3/loss22 = 0.000981242 (* 0.0909091 = 8.92038e-05 loss) | |
| I0510 17:18:45.512639 10926 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0510 17:18:45.512650 10926 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0510 17:18:45.512671 10926 solver.cpp:245] Train net output #149: total_confidence = 4.22128e-05 | |
| I0510 17:18:45.512686 10926 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.00010326 | |
| I0510 17:18:45.512699 10926 sgd_solver.cpp:106] Iteration 22000, lr = 0.001 | |
| I0510 17:18:47.370831 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 32.3994 > 30) by scale factor 0.925942 | |
| I0510 17:20:48.075271 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 63.3835 > 30) by scale factor 0.47331 | |
| I0510 17:20:51.910730 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 34.6087 > 30) by scale factor 0.866835 | |
| I0510 17:21:12.690770 10926 solver.cpp:229] Iteration 22500, loss = 9.66693 | |
| I0510 17:21:12.690831 10926 solver.cpp:245] Train net output #0: loss1/accuracy = 0.133333 | |
| I0510 17:21:12.690860 10926 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.125 | |
| I0510 17:21:12.690886 10926 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0510 17:21:12.690909 10926 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0510 17:21:12.690932 10926 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.5 | |
| I0510 17:21:12.690954 10926 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.375 | |
| I0510 17:21:12.690979 10926 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.5 | |
| I0510 17:21:12.691005 10926 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.625 | |
| I0510 17:21:12.691030 10926 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.75 | |
| I0510 17:21:12.691052 10926 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.875 | |
| I0510 17:21:12.691076 10926 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.875 | |
| I0510 17:21:12.691112 10926 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0510 17:21:12.691134 10926 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0510 17:21:12.691156 10926 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0510 17:21:12.691177 10926 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0510 17:21:12.691198 10926 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0510 17:21:12.691220 10926 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0510 17:21:12.691242 10926 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0510 17:21:12.691263 10926 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0510 17:21:12.691285 10926 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0510 17:21:12.691306 10926 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0510 17:21:12.691328 10926 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0510 17:21:12.691349 10926 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0510 17:21:12.691370 10926 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.767045 | |
| I0510 17:21:12.691395 10926 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.311111 | |
| I0510 17:21:12.691428 10926 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 2.89236 (* 0.3 = 0.867707 loss) | |
| I0510 17:21:12.691467 10926 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 0.836594 (* 0.3 = 0.250978 loss) | |
| I0510 17:21:12.691494 10926 solver.cpp:245] Train net output #27: loss1/loss01 = 2.77568 (* 0.0272727 = 0.0757004 loss) | |
| I0510 17:21:12.691522 10926 solver.cpp:245] Train net output #28: loss1/loss02 = 2.53644 (* 0.0272727 = 0.0691756 loss) | |
| I0510 17:21:12.691548 10926 solver.cpp:245] Train net output #29: loss1/loss03 = 3.3266 (* 0.0272727 = 0.0907253 loss) | |
| I0510 17:21:12.691575 10926 solver.cpp:245] Train net output #30: loss1/loss04 = 1.96912 (* 0.0272727 = 0.0537033 loss) | |
| I0510 17:21:12.691602 10926 solver.cpp:245] Train net output #31: loss1/loss05 = 2.33071 (* 0.0272727 = 0.0635648 loss) | |
| I0510 17:21:12.691628 10926 solver.cpp:245] Train net output #32: loss1/loss06 = 1.6132 (* 0.0272727 = 0.0439962 loss) | |
| I0510 17:21:12.691654 10926 solver.cpp:245] Train net output #33: loss1/loss07 = 1.56715 (* 0.0272727 = 0.0427405 loss) | |
| I0510 17:21:12.691680 10926 solver.cpp:245] Train net output #34: loss1/loss08 = 1.07375 (* 0.0272727 = 0.029284 loss) | |
| I0510 17:21:12.691711 10926 solver.cpp:245] Train net output #35: loss1/loss09 = 0.808537 (* 0.0272727 = 0.022051 loss) | |
| I0510 17:21:12.691738 10926 solver.cpp:245] Train net output #36: loss1/loss10 = 0.614085 (* 0.0272727 = 0.0167478 loss) | |
| I0510 17:21:12.691766 10926 solver.cpp:245] Train net output #37: loss1/loss11 = 0.0136045 (* 0.0272727 = 0.000371032 loss) | |
| I0510 17:21:12.691828 10926 solver.cpp:245] Train net output #38: loss1/loss12 = 0.0135411 (* 0.0272727 = 0.000369302 loss) | |
| I0510 17:21:12.691856 10926 solver.cpp:245] Train net output #39: loss1/loss13 = 0.0050249 (* 0.0272727 = 0.000137043 loss) | |
| I0510 17:21:12.691884 10926 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0064432 (* 0.0272727 = 0.000175724 loss) | |
| I0510 17:21:12.691911 10926 solver.cpp:245] Train net output #41: loss1/loss15 = 0.00853039 (* 0.0272727 = 0.000232647 loss) | |
| I0510 17:21:12.691938 10926 solver.cpp:245] Train net output #42: loss1/loss16 = 0.00626518 (* 0.0272727 = 0.000170869 loss) | |
| I0510 17:21:12.691965 10926 solver.cpp:245] Train net output #43: loss1/loss17 = 0.00253853 (* 0.0272727 = 6.92325e-05 loss) | |
| I0510 17:21:12.691992 10926 solver.cpp:245] Train net output #44: loss1/loss18 = 0.00558199 (* 0.0272727 = 0.000152236 loss) | |
| I0510 17:21:12.692018 10926 solver.cpp:245] Train net output #45: loss1/loss19 = 0.00190414 (* 0.0272727 = 5.19312e-05 loss) | |
| I0510 17:21:12.692046 10926 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00206288 (* 0.0272727 = 5.62603e-05 loss) | |
| I0510 17:21:12.692073 10926 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00111814 (* 0.0272727 = 3.04947e-05 loss) | |
| I0510 17:21:12.692101 10926 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0018075 (* 0.0272727 = 4.92953e-05 loss) | |
| I0510 17:21:12.692123 10926 solver.cpp:245] Train net output #49: loss2/accuracy = 0.155556 | |
| I0510 17:21:12.692150 10926 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.125 | |
| I0510 17:21:12.692173 10926 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0.125 | |
| I0510 17:21:12.692194 10926 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0.125 | |
| I0510 17:21:12.692216 10926 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.375 | |
| I0510 17:21:12.692239 10926 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.375 | |
| I0510 17:21:12.692260 10926 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.5 | |
| I0510 17:21:12.692282 10926 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.625 | |
| I0510 17:21:12.692303 10926 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.75 | |
| I0510 17:21:12.692324 10926 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.875 | |
| I0510 17:21:12.692347 10926 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.875 | |
| I0510 17:21:12.692368 10926 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0510 17:21:12.692389 10926 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0510 17:21:12.692411 10926 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0510 17:21:12.692432 10926 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0510 17:21:12.692453 10926 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0510 17:21:12.692474 10926 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0510 17:21:12.692494 10926 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0510 17:21:12.692517 10926 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0510 17:21:12.692540 10926 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0510 17:21:12.692558 10926 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0510 17:21:12.692580 10926 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0510 17:21:12.692611 10926 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0510 17:21:12.692636 10926 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.778409 | |
| I0510 17:21:12.692658 10926 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.311111 | |
| I0510 17:21:12.692685 10926 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 2.95589 (* 0.3 = 0.886768 loss) | |
| I0510 17:21:12.692711 10926 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 0.833115 (* 0.3 = 0.249935 loss) | |
| I0510 17:21:12.692760 10926 solver.cpp:245] Train net output #76: loss2/loss01 = 2.78277 (* 0.0272727 = 0.0758938 loss) | |
| I0510 17:21:12.692788 10926 solver.cpp:245] Train net output #77: loss2/loss02 = 3.09384 (* 0.0272727 = 0.0843775 loss) | |
| I0510 17:21:12.692814 10926 solver.cpp:245] Train net output #78: loss2/loss03 = 3.02333 (* 0.0272727 = 0.0824544 loss) | |
| I0510 17:21:12.692841 10926 solver.cpp:245] Train net output #79: loss2/loss04 = 2.52138 (* 0.0272727 = 0.0687649 loss) | |
| I0510 17:21:12.692865 10926 solver.cpp:245] Train net output #80: loss2/loss05 = 2.46214 (* 0.0272727 = 0.0671492 loss) | |
| I0510 17:21:12.692893 10926 solver.cpp:245] Train net output #81: loss2/loss06 = 1.64397 (* 0.0272727 = 0.0448356 loss) | |
| I0510 17:21:12.692919 10926 solver.cpp:245] Train net output #82: loss2/loss07 = 1.68205 (* 0.0272727 = 0.045874 loss) | |
| I0510 17:21:12.692945 10926 solver.cpp:245] Train net output #83: loss2/loss08 = 1.12654 (* 0.0272727 = 0.0307238 loss) | |
| I0510 17:21:12.692970 10926 solver.cpp:245] Train net output #84: loss2/loss09 = 0.649987 (* 0.0272727 = 0.0177269 loss) | |
| I0510 17:21:12.692997 10926 solver.cpp:245] Train net output #85: loss2/loss10 = 0.660317 (* 0.0272727 = 0.0180086 loss) | |
| I0510 17:21:12.693024 10926 solver.cpp:245] Train net output #86: loss2/loss11 = 0.0099878 (* 0.0272727 = 0.000272394 loss) | |
| I0510 17:21:12.693050 10926 solver.cpp:245] Train net output #87: loss2/loss12 = 0.0151907 (* 0.0272727 = 0.000414292 loss) | |
| I0510 17:21:12.693078 10926 solver.cpp:245] Train net output #88: loss2/loss13 = 0.0141357 (* 0.0272727 = 0.000385519 loss) | |
| I0510 17:21:12.693104 10926 solver.cpp:245] Train net output #89: loss2/loss14 = 0.00475208 (* 0.0272727 = 0.000129602 loss) | |
| I0510 17:21:12.693147 10926 solver.cpp:245] Train net output #90: loss2/loss15 = 0.00492765 (* 0.0272727 = 0.000134391 loss) | |
| I0510 17:21:12.693179 10926 solver.cpp:245] Train net output #91: loss2/loss16 = 0.00244329 (* 0.0272727 = 6.66351e-05 loss) | |
| I0510 17:21:12.693210 10926 solver.cpp:245] Train net output #92: loss2/loss17 = 0.00450313 (* 0.0272727 = 0.000122813 loss) | |
| I0510 17:21:12.693238 10926 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0017603 (* 0.0272727 = 4.80081e-05 loss) | |
| I0510 17:21:12.693265 10926 solver.cpp:245] Train net output #94: loss2/loss19 = 0.0028513 (* 0.0272727 = 7.77628e-05 loss) | |
| I0510 17:21:12.693292 10926 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00340637 (* 0.0272727 = 9.2901e-05 loss) | |
| I0510 17:21:12.693318 10926 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00225897 (* 0.0272727 = 6.16082e-05 loss) | |
| I0510 17:21:12.693346 10926 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0010441 (* 0.0272727 = 2.84754e-05 loss) | |
| I0510 17:21:12.693367 10926 solver.cpp:245] Train net output #98: loss3/accuracy = 0.111111 | |
| I0510 17:21:12.693389 10926 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.375 | |
| I0510 17:21:12.693413 10926 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.25 | |
| I0510 17:21:12.693434 10926 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0.25 | |
| I0510 17:21:12.693454 10926 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.375 | |
| I0510 17:21:12.693476 10926 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.375 | |
| I0510 17:21:12.693497 10926 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.5 | |
| I0510 17:21:12.693517 10926 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.625 | |
| I0510 17:21:12.693539 10926 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.75 | |
| I0510 17:21:12.693560 10926 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.875 | |
| I0510 17:21:12.693581 10926 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.875 | |
| I0510 17:21:12.693603 10926 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0510 17:21:12.693624 10926 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0510 17:21:12.693645 10926 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0510 17:21:12.693683 10926 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0510 17:21:12.693707 10926 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0510 17:21:12.693728 10926 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0510 17:21:12.693749 10926 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0510 17:21:12.693770 10926 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0510 17:21:12.693791 10926 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0510 17:21:12.693815 10926 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0510 17:21:12.693836 10926 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0510 17:21:12.693858 10926 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0510 17:21:12.693879 10926 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.767045 | |
| I0510 17:21:12.693902 10926 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.333333 | |
| I0510 17:21:12.693925 10926 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 2.94296 (* 1 = 2.94296 loss) | |
| I0510 17:21:12.693951 10926 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 0.804214 (* 1 = 0.804214 loss) | |
| I0510 17:21:12.693976 10926 solver.cpp:245] Train net output #125: loss3/loss01 = 2.34608 (* 0.0909091 = 0.21328 loss) | |
| I0510 17:21:12.694003 10926 solver.cpp:245] Train net output #126: loss3/loss02 = 2.57178 (* 0.0909091 = 0.233798 loss) | |
| I0510 17:21:12.694030 10926 solver.cpp:245] Train net output #127: loss3/loss03 = 3.01723 (* 0.0909091 = 0.274293 loss) | |
| I0510 17:21:12.694056 10926 solver.cpp:245] Train net output #128: loss3/loss04 = 1.84073 (* 0.0909091 = 0.167339 loss) | |
| I0510 17:21:12.694080 10926 solver.cpp:245] Train net output #129: loss3/loss05 = 1.82231 (* 0.0909091 = 0.165665 loss) | |
| I0510 17:21:12.694106 10926 solver.cpp:245] Train net output #130: loss3/loss06 = 1.59574 (* 0.0909091 = 0.145067 loss) | |
| I0510 17:21:12.694131 10926 solver.cpp:245] Train net output #131: loss3/loss07 = 1.13987 (* 0.0909091 = 0.103625 loss) | |
| I0510 17:21:12.694156 10926 solver.cpp:245] Train net output #132: loss3/loss08 = 1.12598 (* 0.0909091 = 0.102362 loss) | |
| I0510 17:21:12.694182 10926 solver.cpp:245] Train net output #133: loss3/loss09 = 0.712394 (* 0.0909091 = 0.0647631 loss) | |
| I0510 17:21:12.694208 10926 solver.cpp:245] Train net output #134: loss3/loss10 = 0.559602 (* 0.0909091 = 0.0508729 loss) | |
| I0510 17:21:12.694236 10926 solver.cpp:245] Train net output #135: loss3/loss11 = 0.00551734 (* 0.0909091 = 0.000501577 loss) | |
| I0510 17:21:12.694264 10926 solver.cpp:245] Train net output #136: loss3/loss12 = 0.00341952 (* 0.0909091 = 0.000310866 loss) | |
| I0510 17:21:12.694290 10926 solver.cpp:245] Train net output #137: loss3/loss13 = 0.0030147 (* 0.0909091 = 0.000274064 loss) | |
| I0510 17:21:12.694316 10926 solver.cpp:245] Train net output #138: loss3/loss14 = 0.0027686 (* 0.0909091 = 0.000251691 loss) | |
| I0510 17:21:12.694344 10926 solver.cpp:245] Train net output #139: loss3/loss15 = 0.00162471 (* 0.0909091 = 0.000147701 loss) | |
| I0510 17:21:12.694368 10926 solver.cpp:245] Train net output #140: loss3/loss16 = 0.00132908 (* 0.0909091 = 0.000120826 loss) | |
| I0510 17:21:12.694394 10926 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00108441 (* 0.0909091 = 9.85825e-05 loss) | |
| I0510 17:21:12.694421 10926 solver.cpp:245] Train net output #142: loss3/loss18 = 0.000951044 (* 0.0909091 = 8.64586e-05 loss) | |
| I0510 17:21:12.694447 10926 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00160625 (* 0.0909091 = 0.000146023 loss) | |
| I0510 17:21:12.694473 10926 solver.cpp:245] Train net output #144: loss3/loss20 = 0.000854793 (* 0.0909091 = 7.77085e-05 loss) | |
| I0510 17:21:12.694499 10926 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00077443 (* 0.0909091 = 7.04027e-05 loss) | |
| I0510 17:21:12.694525 10926 solver.cpp:245] Train net output #146: loss3/loss22 = 0.000874352 (* 0.0909091 = 7.94865e-05 loss) | |
| I0510 17:21:12.694561 10926 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0510 17:21:12.694584 10926 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0.125 | |
| I0510 17:21:12.694605 10926 solver.cpp:245] Train net output #149: total_confidence = 0.000361461 | |
| I0510 17:21:12.694628 10926 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.000761178 | |
| I0510 17:21:12.694650 10926 sgd_solver.cpp:106] Iteration 22500, lr = 0.001 | |
| I0510 17:21:33.684834 10926 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 61.438 > 30) by scale factor 0.488297 | |
| I0510 17:23:39.963302 10926 solver.cpp:229] Iteration 23000, loss = 9.65796 | |
| I0510 17:23:39.963825 10926 solver.cpp:245] Train net output #0: loss1/accuracy = 0.136364 | |
| I0510 17:23:39.963845 10926 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.25 | |
| I0510 17:23:39.963860 10926 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.125 | |
| I0510 17:23:39.963872 10926 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.125 | |
| I0510 17:23:39.963886 10926 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.125 | |
| I0510 17:23:39.963897 10926 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.25 | |
| I0510 17:23:39.963910 10926 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.625 | |
| I0510 17:23:39.963923 10926 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.875 | |
| I0510 17:23:39.963935 10926 solver.cpp:24 |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment