magicpoint_synth

2018-12-21 19:38:48.457668: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1484] Adding visible gpu devices: 0
2018-12-21 19:38:48.457719: I tensorflow/core/common_runtime/gpu/gpu_device.cc:965] Device interconnect StreamExecutor with strength 1 edge matrix:
2018-12-21 19:38:48.457728: I tensorflow/core/common_runtime/gpu/gpu_device.cc:971]      0 
2018-12-21 19:38:48.457734: I tensorflow/core/common_runtime/gpu/gpu_device.cc:984] 0:   N 
2018-12-21 19:38:48.457878: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1097] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 10410 MB memory) -> physical GPU (device: 0, name: GeForce GTX 1080 Ti, pci bus id: 0000:03:00.0, compute capability: 6.1)
[12/21/2018 19:38:54 INFO] Start training
[12/21/2018 19:39:44 INFO] Iter    0: loss 4.6639, precision 0.0006, recall 0.0562
[12/21/2018 19:49:00 INFO] Iter 1000: loss 1.4334, precision 0.0027, recall 0.2551
[12/21/2018 19:57:41 INFO] Iter 2000: loss 0.5045, precision 0.0033, recall 0.3159
[12/21/2018 20:05:51 INFO] Iter 3000: loss 0.2442, precision 0.0038, recall 0.3523
[12/21/2018 20:13:57 INFO] Iter 4000: loss 0.1585, precision 0.0047, recall 0.4260
[12/21/2018 20:21:38 INFO] Iter 5000: loss 0.1168, precision 0.1120, recall 0.4207
[12/21/2018 20:29:20 INFO] Iter 6000: loss 0.1082, precision 0.1584, recall 0.4482
[12/21/2018 20:37:04 INFO] Iter 7000: loss 0.0972, precision 0.1880, recall 0.4554
[12/21/2018 20:44:45 INFO] Iter 8000: loss 0.0894, precision 0.1850, recall 0.3925
[12/21/2018 20:52:28 INFO] Iter 9000: loss 0.0843, precision 0.2616, recall 0.4619
[12/21/2018 21:00:12 INFO] Iter 10000: loss 0.0817, precision 0.2335, recall 0.4585
[12/21/2018 21:07:57 INFO] Iter 11000: loss 0.0829, precision 0.2466, recall 0.4474
[12/21/2018 21:15:35 INFO] Iter 12000: loss 0.0806, precision 0.3073, recall 0.5054
[12/21/2018 21:23:16 INFO] Iter 13000: loss 0.0965, precision 0.2495, recall 0.4467
[12/21/2018 21:30:57 INFO] Iter 14000: loss 0.0875, precision 0.1902, recall 0.4012
[12/21/2018 21:38:40 INFO] Iter 15000: loss 0.0848, precision 0.3155, recall 0.4994
[12/21/2018 21:46:21 INFO] Iter 16000: loss 0.0635, precision 0.3362, recall 0.5149
[12/21/2018 21:54:02 INFO] Iter 17000: loss 0.0750, precision 0.3143, recall 0.4868
[12/21/2018 22:01:43 INFO] Iter 18000: loss 0.0636, precision 0.3119, recall 0.4911
[12/21/2018 22:09:26 INFO] Iter 19000: loss 0.0729, precision 0.2722, recall 0.4865
[12/21/2018 22:17:08 INFO] Iter 20000: loss 0.0788, precision 0.3344, recall 0.5279
[12/21/2018 22:24:51 INFO] Iter 21000: loss 0.1236, precision 0.3063, recall 0.5041
[12/21/2018 22:32:31 INFO] Iter 22000: loss 0.0839, precision 0.2948, recall 0.4696
[12/21/2018 22:40:10 INFO] Iter 23000: loss 0.0682, precision 0.3153, recall 0.4693
[12/21/2018 22:47:50 INFO] Iter 24000: loss 0.0887, precision 0.3171, recall 0.4833
[12/21/2018 22:55:31 INFO] Iter 25000: loss 0.0593, precision 0.3057, recall 0.4996
[12/21/2018 23:03:13 INFO] Iter 26000: loss 0.0793, precision 0.2995, recall 0.4956
[12/21/2018 23:11:16 INFO] Iter 27000: loss 0.0762, precision 0.3343, recall 0.4988
[12/21/2018 23:19:26 INFO] Iter 28000: loss 0.0555, precision 0.3620, recall 0.5488
[12/21/2018 23:27:10 INFO] Iter 29000: loss 0.0742, precision 0.3610, recall 0.5137
[12/21/2018 23:34:53 INFO] Iter 30000: loss 0.0811, precision 0.3259, recall 0.4887
[12/21/2018 23:42:44 INFO] Iter 31000: loss 0.0679, precision 0.3731, recall 0.5374
[12/21/2018 23:50:25 INFO] Iter 32000: loss 0.0747, precision 0.3257, recall 0.5183
[12/21/2018 23:58:07 INFO] Iter 33000: loss 0.0717, precision 0.3268, recall 0.5266
[12/22/2018 00:05:50 INFO] Iter 34000: loss 0.0746, precision 0.3804, recall 0.5448
[12/22/2018 00:13:37 INFO] Iter 35000: loss 0.0795, precision 0.3485, recall 0.5273
[12/22/2018 00:21:20 INFO] Iter 36000: loss 0.0938, precision 0.3073, recall 0.4645
[12/22/2018 00:29:04 INFO] Iter 37000: loss 0.0812, precision 0.3767, recall 0.5511
[12/22/2018 00:36:44 INFO] Iter 38000: loss 0.0658, precision 0.3489, recall 0.5335
[12/22/2018 00:44:27 INFO] Iter 39000: loss 0.0787, precision 0.3604, recall 0.5467
[12/22/2018 00:52:13 INFO] Iter 40000: loss 0.0674, precision 0.3521, recall 0.5026
[12/22/2018 00:59:53 INFO] Iter 41000: loss 0.0623, precision 0.3790, recall 0.5476
[12/22/2018 01:07:34 INFO] Iter 42000: loss 0.0721, precision 0.3694, recall 0.5225
[12/22/2018 01:15:13 INFO] Iter 43000: loss 0.0713, precision 0.3185, recall 0.4988
[12/22/2018 01:22:52 INFO] Iter 44000: loss 0.0770, precision 0.3354, recall 0.4814
[12/22/2018 01:30:30 INFO] Iter 45000: loss 0.0669, precision 0.3682, recall 0.5221
[12/22/2018 01:38:09 INFO] Iter 46000: loss 0.0827, precision 0.3462, recall 0.5076
[12/22/2018 01:45:48 INFO] Iter 47000: loss 0.0761, precision 0.3415, recall 0.4967
[12/22/2018 01:53:29 INFO] Iter 48000: loss 0.0802, precision 0.3148, recall 0.4953
[12/22/2018 02:01:09 INFO] Iter 49000: loss 0.0692, precision 0.3260, recall 0.4843
[12/22/2018 02:08:48 INFO] Iter 50000: loss 0.0620, precision 0.3046, recall 0.4688
[12/22/2018 02:16:30 INFO] Iter 51000: loss 0.0603, precision 0.3809, recall 0.5539
[12/22/2018 02:24:09 INFO] Iter 52000: loss 0.0697, precision 0.2922, recall 0.4656
[12/22/2018 02:31:49 INFO] Iter 53000: loss 0.0755, precision 0.3161, recall 0.4817
[12/22/2018 02:39:29 INFO] Iter 54000: loss 0.0933, precision 0.3499, recall 0.5251
[12/22/2018 02:47:07 INFO] Iter 55000: loss 0.0811, precision 0.3785, recall 0.5455
[12/22/2018 02:54:47 INFO] Iter 56000: loss 0.0749, precision 0.3863, recall 0.5623
[12/22/2018 03:02:27 INFO] Iter 57000: loss 0.0777, precision 0.3365, recall 0.5119
[12/22/2018 03:10:06 INFO] Iter 58000: loss 0.0646, precision 0.3780, recall 0.5493
[12/22/2018 03:17:46 INFO] Iter 59000: loss 0.0770, precision 0.3746, recall 0.5544
[12/22/2018 03:25:25 INFO] Iter 60000: loss 0.0720, precision 0.3472, recall 0.4917
[12/22/2018 03:33:05 INFO] Iter 61000: loss 0.0551, precision 0.2775, recall 0.4428
[12/22/2018 03:40:44 INFO] Iter 62000: loss 0.0796, precision 0.2903, recall 0.4595
[12/22/2018 03:48:24 INFO] Iter 63000: loss 0.0745, precision 0.3340, recall 0.4665
[12/22/2018 03:56:04 INFO] Iter 64000: loss 0.0635, precision 0.3272, recall 0.4897
[12/22/2018 04:03:42 INFO] Iter 65000: loss 0.0713, precision 0.3614, recall 0.5414
[12/22/2018 04:11:21 INFO] Iter 66000: loss 0.0750, precision 0.3072, recall 0.4863
[12/22/2018 04:19:03 INFO] Iter 67000: loss 0.0713, precision 0.3552, recall 0.5207
[12/22/2018 04:26:42 INFO] Iter 68000: loss 0.0809, precision 0.3683, recall 0.5369
[12/22/2018 04:34:23 INFO] Iter 69000: loss 0.0829, precision 0.2953, recall 0.4507
[12/22/2018 04:42:02 INFO] Iter 70000: loss 0.0661, precision 0.3831, recall 0.5450
[12/22/2018 04:49:41 INFO] Iter 71000: loss 0.0695, precision 0.2914, recall 0.4556
[12/22/2018 04:57:23 INFO] Iter 72000: loss 0.0666, precision 0.3320, recall 0.4827
[12/22/2018 05:05:01 INFO] Iter 73000: loss 0.0637, precision 0.3391, recall 0.4978
[12/22/2018 05:12:42 INFO] Iter 74000: loss 0.0629, precision 0.2586, recall 0.3961
[12/22/2018 05:20:18 INFO] Iter 75000: loss 0.0622, precision 0.3234, recall 0.4698
[12/22/2018 05:27:59 INFO] Iter 76000: loss 0.0715, precision 0.3017, recall 0.4908
[12/22/2018 05:35:42 INFO] Iter 77000: loss 0.0741, precision 0.4034, recall 0.5765
[12/22/2018 05:43:20 INFO] Iter 78000: loss 0.0659, precision 0.3396, recall 0.4914
[12/22/2018 05:51:00 INFO] Iter 79000: loss 0.0839, precision 0.3127, recall 0.4758
[12/22/2018 05:58:40 INFO] Iter 80000: loss 0.0731, precision 0.3050, recall 0.4605
[12/22/2018 06:06:18 INFO] Iter 81000: loss 0.0757, precision 0.3827, recall 0.5497
[12/22/2018 06:13:56 INFO] Iter 82000: loss 0.0679, precision 0.3067, recall 0.4548
[12/22/2018 06:21:36 INFO] Iter 83000: loss 0.0663, precision 0.3878, recall 0.5599
[12/22/2018 06:29:16 INFO] Iter 84000: loss 0.0558, precision 0.3671, recall 0.5206
[12/22/2018 06:36:54 INFO] Iter 85000: loss 0.0586, precision 0.3260, recall 0.4856
[12/22/2018 06:44:32 INFO] Iter 86000: loss 0.0630, precision 0.3894, recall 0.5405
[12/22/2018 06:52:12 INFO] Iter 87000: loss 0.0878, precision 0.3162, recall 0.4461
[12/22/2018 06:59:50 INFO] Iter 88000: loss 0.0826, precision 0.3790, recall 0.5182
[12/22/2018 07:07:28 INFO] Iter 89000: loss 0.0661, precision 0.3794, recall 0.5333
[12/22/2018 07:15:07 INFO] Iter 90000: loss 0.0782, precision 0.4081, recall 0.5573
[12/22/2018 07:22:47 INFO] Iter 91000: loss 0.0636, precision 0.3588, recall 0.5307
[12/22/2018 07:30:28 INFO] Iter 92000: loss 0.0730, precision 0.3883, recall 0.5683
[12/22/2018 07:38:07 INFO] Iter 93000: loss 0.0648, precision 0.3703, recall 0.5098
[12/22/2018 07:45:46 INFO] Iter 94000: loss 0.0538, precision 0.3766, recall 0.5552
[12/22/2018 07:53:24 INFO] Iter 95000: loss 0.0681, precision 0.3642, recall 0.5398
[12/22/2018 08:01:05 INFO] Iter 96000: loss 0.0744, precision 0.3713, recall 0.5148
[12/22/2018 08:08:45 INFO] Iter 97000: loss 0.0592, precision 0.3295, recall 0.4776
[12/22/2018 08:16:25 INFO] Iter 98000: loss 0.0728, precision 0.3933, recall 0.5687
[12/22/2018 08:24:07 INFO] Iter 99000: loss 0.0636, precision 0.3524, recall 0.5101
[12/22/2018 08:31:51 INFO] Iter 100000: loss 0.0783, precision 0.3482, recall 0.5149
[12/22/2018 08:39:31 INFO] Iter 101000: loss 0.0674, precision 0.3531, recall 0.4867
[12/22/2018 08:47:10 INFO] Iter 102000: loss 0.0792, precision 0.4068, recall 0.5684
[12/22/2018 08:54:48 INFO] Iter 103000: loss 0.0684, precision 0.3628, recall 0.5240
[12/22/2018 09:02:28 INFO] Iter 104000: loss 0.0624, precision 0.3470, recall 0.4951
[12/22/2018 09:10:07 INFO] Iter 105000: loss 0.0737, precision 0.3934, recall 0.5517
[12/22/2018 09:17:48 INFO] Iter 106000: loss 0.0770, precision 0.3302, recall 0.4696
[12/22/2018 09:25:29 INFO] Iter 107000: loss 0.0606, precision 0.3594, recall 0.5118
[12/22/2018 09:33:06 INFO] Iter 108000: loss 0.0616, precision 0.3831, recall 0.5432
[12/22/2018 09:40:47 INFO] Iter 109000: loss 0.0596, precision 0.3469, recall 0.4999
[12/22/2018 09:48:23 INFO] Iter 110000: loss 0.0792, precision 0.3516, recall 0.5127
[12/22/2018 09:56:03 INFO] Iter 111000: loss 0.0580, precision 0.3184, recall 0.4701
[12/22/2018 10:03:45 INFO] Iter 112000: loss 0.0666, precision 0.2876, recall 0.3994
[12/22/2018 10:11:20 INFO] Iter 113000: loss 0.0674, precision 0.3497, recall 0.5026
[12/22/2018 10:19:00 INFO] Iter 114000: loss 0.0966, precision 0.4128, recall 0.5701
[12/22/2018 10:26:40 INFO] Iter 115000: loss 0.0580, precision 0.3990, recall 0.5406
[12/22/2018 10:34:21 INFO] Iter 116000: loss 0.0734, precision 0.3202, recall 0.4633
[12/22/2018 10:42:02 INFO] Iter 117000: loss 0.0586, precision 0.3192, recall 0.4605
[12/22/2018 10:49:41 INFO] Iter 118000: loss 0.0564, precision 0.3153, recall 0.4471
[12/22/2018 10:57:19 INFO] Iter 119000: loss 0.0569, precision 0.3839, recall 0.5337
[12/22/2018 11:04:58 INFO] Iter 120000: loss 0.0861, precision 0.3208, recall 0.4657
[12/22/2018 11:12:37 INFO] Iter 121000: loss 0.0745, precision 0.3242, recall 0.4688
[12/22/2018 11:20:18 INFO] Iter 122000: loss 0.0734, precision 0.3530, recall 0.5147
[12/22/2018 11:27:56 INFO] Iter 123000: loss 0.0753, precision 0.3177, recall 0.4615
[12/22/2018 11:35:36 INFO] Iter 124000: loss 0.0561, precision 0.3155, recall 0.4652
[12/22/2018 11:43:15 INFO] Iter 125000: loss 0.0718, precision 0.2952, recall 0.4441
[12/22/2018 11:50:54 INFO] Iter 126000: loss 0.0660, precision 0.2503, recall 0.3678
[12/22/2018 11:58:35 INFO] Iter 127000: loss 0.0497, precision 0.3192, recall 0.4537
[12/22/2018 12:06:16 INFO] Iter 128000: loss 0.0657, precision 0.3896, recall 0.5521
[12/22/2018 12:13:57 INFO] Iter 129000: loss 0.0731, precision 0.3790, recall 0.5276
[12/22/2018 12:21:36 INFO] Iter 130000: loss 0.0815, precision 0.2897, recall 0.4210
[12/22/2018 12:29:16 INFO] Iter 131000: loss 0.0701, precision 0.3577, recall 0.4942
[12/22/2018 12:36:58 INFO] Iter 132000: loss 0.0657, precision 0.3522, recall 0.5008
[12/22/2018 12:44:36 INFO] Iter 133000: loss 0.0790, precision 0.2987, recall 0.4375
[12/22/2018 12:52:17 INFO] Iter 134000: loss 0.0795, precision 0.3408, recall 0.4963
[12/22/2018 12:59:59 INFO] Iter 135000: loss 0.0714, precision 0.3432, recall 0.5028
[12/22/2018 13:07:37 INFO] Iter 136000: loss 0.0603, precision 0.3718, recall 0.5150
[12/22/2018 13:15:16 INFO] Iter 137000: loss 0.0709, precision 0.2820, recall 0.4029
[12/22/2018 13:22:57 INFO] Iter 138000: loss 0.0637, precision 0.4112, recall 0.5649
[12/22/2018 13:30:38 INFO] Iter 139000: loss 0.0591, precision 0.4046, recall 0.5616
[12/22/2018 13:38:16 INFO] Iter 140000: loss 0.0643, precision 0.3390, recall 0.4908
[12/22/2018 13:45:55 INFO] Iter 141000: loss 0.0569, precision 0.3363, recall 0.4757
[12/22/2018 13:53:36 INFO] Iter 142000: loss 0.0635, precision 0.3130, recall 0.4461
[12/22/2018 14:01:15 INFO] Iter 143000: loss 0.0571, precision 0.2144, recall 0.3095
[12/22/2018 14:08:55 INFO] Iter 144000: loss 0.0619, precision 0.3131, recall 0.4353
[12/22/2018 14:16:32 INFO] Iter 145000: loss 0.0696, precision 0.4133, recall 0.5539
[12/22/2018 14:24:11 INFO] Iter 146000: loss 0.0772, precision 0.4262, recall 0.5827
[12/22/2018 14:31:51 INFO] Iter 147000: loss 0.0687, precision 0.2568, recall 0.3719
[12/22/2018 14:39:31 INFO] Iter 148000: loss 0.0687, precision 0.3482, recall 0.4776
[12/22/2018 14:47:11 INFO] Iter 149000: loss 0.0570, precision 0.4090, recall 0.5744
[12/22/2018 14:54:49 INFO] Iter 150000: loss 0.0473, precision 0.3674, recall 0.5014
[12/22/2018 15:02:27 INFO] Iter 151000: loss 0.0691, precision 0.3667, recall 0.5019
[12/22/2018 15:10:06 INFO] Iter 152000: loss 0.0584, precision 0.4256, recall 0.5775
[12/22/2018 15:17:44 INFO] Iter 153000: loss 0.0923, precision 0.3880, recall 0.5541
[12/22/2018 15:25:23 INFO] Iter 154000: loss 0.0728, precision 0.3628, recall 0.5200
[12/22/2018 15:33:02 INFO] Iter 155000: loss 0.0661, precision 0.3389, recall 0.4817
[12/22/2018 15:40:41 INFO] Iter 156000: loss 0.0775, precision 0.4143, recall 0.5749
[12/22/2018 15:48:21 INFO] Iter 157000: loss 0.0562, precision 0.3743, recall 0.5172
[12/22/2018 15:56:02 INFO] Iter 158000: loss 0.0729, precision 0.3378, recall 0.4925
[12/22/2018 16:03:38 INFO] Iter 159000: loss 0.0617, precision 0.2945, recall 0.4253
[12/22/2018 16:11:16 INFO] Iter 160000: loss 0.0494, precision 0.3138, recall 0.4437
[12/22/2018 16:18:55 INFO] Iter 161000: loss 0.0705, precision 0.3357, recall 0.4745
[12/22/2018 16:26:35 INFO] Iter 162000: loss 0.0735, precision 0.3354, recall 0.4993
[12/22/2018 16:34:14 INFO] Iter 163000: loss 0.0713, precision 0.3609, recall 0.5171
[12/22/2018 16:41:54 INFO] Iter 164000: loss 0.0730, precision 0.3859, recall 0.5299
[12/22/2018 16:49:35 INFO] Iter 165000: loss 0.0618, precision 0.3036, recall 0.4413
[12/22/2018 16:57:17 INFO] Iter 166000: loss 0.0725, precision 0.3506, recall 0.5180
[12/22/2018 17:04:55 INFO] Iter 167000: loss 0.0697, precision 0.3613, recall 0.5097
[12/22/2018 17:12:35 INFO] Iter 168000: loss 0.0740, precision 0.3735, recall 0.5302
[12/22/2018 17:20:16 INFO] Iter 169000: loss 0.0579, precision 0.3742, recall 0.5193
[12/22/2018 17:27:53 INFO] Iter 170000: loss 0.0668, precision 0.3864, recall 0.5132
[12/22/2018 17:35:31 INFO] Iter 171000: loss 0.0626, precision 0.3845, recall 0.5216
[12/22/2018 17:43:11 INFO] Iter 172000: loss 0.0646, precision 0.3606, recall 0.5181
[12/22/2018 17:50:54 INFO] Iter 173000: loss 0.0629, precision 0.3512, recall 0.4917
[12/22/2018 17:58:35 INFO] Iter 174000: loss 0.0506, precision 0.3613, recall 0.4966
[12/22/2018 18:06:14 INFO] Iter 175000: loss 0.0634, precision 0.3277, recall 0.4651
[12/22/2018 18:13:54 INFO] Iter 176000: loss 0.0686, precision 0.3693, recall 0.5254
[12/22/2018 18:21:38 INFO] Iter 177000: loss 0.0612, precision 0.3516, recall 0.4876
[12/22/2018 18:29:21 INFO] Iter 178000: loss 0.0834, precision 0.3380, recall 0.4870
[12/22/2018 18:37:08 INFO] Iter 179000: loss 0.0731, precision 0.3461, recall 0.4825
[12/22/2018 18:44:54 INFO] Iter 180000: loss 0.0778, precision 0.3554, recall 0.5148
[12/22/2018 18:52:59 INFO] Iter 181000: loss 0.0617, precision 0.4033, recall 0.5473
[12/22/2018 19:01:13 INFO] Iter 182000: loss 0.0603, precision 0.3892, recall 0.5519
[12/22/2018 19:08:55 INFO] Iter 183000: loss 0.0532, precision 0.3479, recall 0.4771
[12/22/2018 19:16:37 INFO] Iter 184000: loss 0.0608, precision 0.3619, recall 0.5044
[12/22/2018 19:24:19 INFO] Iter 185000: loss 0.0604, precision 0.3086, recall 0.4252
[12/22/2018 19:31:58 INFO] Iter 186000: loss 0.0785, precision 0.3974, recall 0.5342
[12/22/2018 19:39:38 INFO] Iter 187000: loss 0.0876, precision 0.3626, recall 0.5151
[12/22/2018 19:47:19 INFO] Iter 188000: loss 0.0645, precision 0.2971, recall 0.4144
[12/22/2018 19:54:59 INFO] Iter 189000: loss 0.0731, precision 0.3955, recall 0.5441
[12/22/2018 20:02:39 INFO] Iter 190000: loss 0.0556, precision 0.4046, recall 0.5605
[12/22/2018 20:10:19 INFO] Iter 191000: loss 0.0751, precision 0.3219, recall 0.4604
[12/22/2018 20:18:00 INFO] Iter 192000: loss 0.0581, precision 0.4066, recall 0.5343
[12/22/2018 20:25:40 INFO] Iter 193000: loss 0.0569, precision 0.3854, recall 0.5272
[12/22/2018 20:33:20 INFO] Iter 194000: loss 0.0673, precision 0.3525, recall 0.4975
[12/22/2018 20:41:03 INFO] Iter 195000: loss 0.0734, precision 0.3533, recall 0.4943
[12/22/2018 20:48:43 INFO] Iter 196000: loss 0.0696, precision 0.3671, recall 0.5302
[12/22/2018 20:56:23 INFO] Iter 197000: loss 0.0837, precision 0.3952, recall 0.5621
[12/22/2018 21:04:05 INFO] Iter 198000: loss 0.0750, precision 0.3811, recall 0.5405
[12/22/2018 21:11:46 INFO] Iter 199000: loss 0.0639, precision 0.3580, recall 0.5218
[12/22/2018 21:19:23 INFO] Training finished
[12/22/2018 21:19:23 INFO] Saving checkpoint for iteration #200000