Baseline log

baseline.log

Loading data from 'data/small-train.t7'...	
 * vocabulary size: source = 50004; target = 50004	
 * additional features: source = 0; target = 0	
 * maximum sequence length: source = 50; target = 51	
 * number of training sentences: 100000	
 * maximum batch size: 64	
Building model...	
 * using input feeding	
Initializing parameters...	
 * number of parameters: 84814004	
Preparing memory optimization...	
 * sharing 69% of output/gradInput tensors memory between clones	
Start training...	
	
Epoch 1 ; Iteration 50/1587 ; Learning rate 1.0000 ; Source tokens/s 732 ; Perplexity 615210.38	
Epoch 1 ; Iteration 100/1587 ; Learning rate 1.0000 ; Source tokens/s 1069 ; Perplexity 105114.37	
Epoch 1 ; Iteration 150/1587 ; Learning rate 1.0000 ; Source tokens/s 1227 ; Perplexity 38544.59	
Epoch 1 ; Iteration 200/1587 ; Learning rate 1.0000 ; Source tokens/s 1371 ; Perplexity 18725.93	
Epoch 1 ; Iteration 250/1587 ; Learning rate 1.0000 ; Source tokens/s 1455 ; Perplexity 11406.14	
Epoch 1 ; Iteration 300/1587 ; Learning rate 1.0000 ; Source tokens/s 1514 ; Perplexity 8083.89	
Epoch 1 ; Iteration 350/1587 ; Learning rate 1.0000 ; Source tokens/s 1543 ; Perplexity 6289.71	
Epoch 1 ; Iteration 400/1587 ; Learning rate 1.0000 ; Source tokens/s 1562 ; Perplexity 5138.64	
Epoch 1 ; Iteration 450/1587 ; Learning rate 1.0000 ; Source tokens/s 1579 ; Perplexity 4310.08	
Epoch 1 ; Iteration 500/1587 ; Learning rate 1.0000 ; Source tokens/s 1615 ; Perplexity 3626.10	
Epoch 1 ; Iteration 550/1587 ; Learning rate 1.0000 ; Source tokens/s 1632 ; Perplexity 3157.90	
Epoch 1 ; Iteration 600/1587 ; Learning rate 1.0000 ; Source tokens/s 1653 ; Perplexity 2777.50	
Epoch 1 ; Iteration 650/1587 ; Learning rate 1.0000 ; Source tokens/s 1659 ; Perplexity 2503.01	
Epoch 1 ; Iteration 700/1587 ; Learning rate 1.0000 ; Source tokens/s 1670 ; Perplexity 2271.96	
Epoch 1 ; Iteration 750/1587 ; Learning rate 1.0000 ; Source tokens/s 1680 ; Perplexity 2080.22	
Epoch 1 ; Iteration 800/1587 ; Learning rate 1.0000 ; Source tokens/s 1693 ; Perplexity 1912.76	
Epoch 1 ; Iteration 850/1587 ; Learning rate 1.0000 ; Source tokens/s 1696 ; Perplexity 1781.79	
Epoch 1 ; Iteration 900/1587 ; Learning rate 1.0000 ; Source tokens/s 1703 ; Perplexity 1661.91	
Epoch 1 ; Iteration 950/1587 ; Learning rate 1.0000 ; Source tokens/s 1700 ; Perplexity 1567.93	
Epoch 1 ; Iteration 1000/1587 ; Learning rate 1.0000 ; Source tokens/s 1712 ; Perplexity 1471.73	
Epoch 1 ; Iteration 1050/1587 ; Learning rate 1.0000 ; Source tokens/s 1715 ; Perplexity 1389.92	
Epoch 1 ; Iteration 1100/1587 ; Learning rate 1.0000 ; Source tokens/s 1723 ; Perplexity 1314.89	
Epoch 1 ; Iteration 1150/1587 ; Learning rate 1.0000 ; Source tokens/s 1726 ; Perplexity 1250.81	
Epoch 1 ; Iteration 1200/1587 ; Learning rate 1.0000 ; Source tokens/s 1729 ; Perplexity 1194.76	
Epoch 1 ; Iteration 1250/1587 ; Learning rate 1.0000 ; Source tokens/s 1731 ; Perplexity 1144.58	
Epoch 1 ; Iteration 1300/1587 ; Learning rate 1.0000 ; Source tokens/s 1740 ; Perplexity 1094.53	
Epoch 1 ; Iteration 1350/1587 ; Learning rate 1.0000 ; Source tokens/s 1743 ; Perplexity 1048.14	
Epoch 1 ; Iteration 1400/1587 ; Learning rate 1.0000 ; Source tokens/s 1745 ; Perplexity 1008.05	
Epoch 1 ; Iteration 1450/1587 ; Learning rate 1.0000 ; Source tokens/s 1751 ; Perplexity 967.57	
Epoch 1 ; Iteration 1500/1587 ; Learning rate 1.0000 ; Source tokens/s 1755 ; Perplexity 932.41	
Epoch 1 ; Iteration 1550/1587 ; Learning rate 1.0000 ; Source tokens/s 1759 ; Perplexity 898.71	
Validation perplexity: 275.10064334583	
Saving checkpoint to 'models/baseline_epoch1_275.10.t7'...	
	
Epoch 2 ; Iteration 50/1587 ; Learning rate 1.0000 ; Source tokens/s 1788 ; Perplexity 270.24	
Epoch 2 ; Iteration 100/1587 ; Learning rate 1.0000 ; Source tokens/s 1855 ; Perplexity 274.90	
Epoch 2 ; Iteration 150/1587 ; Learning rate 1.0000 ; Source tokens/s 1824 ; Perplexity 275.08	
Epoch 2 ; Iteration 200/1587 ; Learning rate 1.0000 ; Source tokens/s 1805 ; Perplexity 271.69	
Epoch 2 ; Iteration 250/1587 ; Learning rate 1.0000 ; Source tokens/s 1818 ; Perplexity 267.60	
Epoch 2 ; Iteration 300/1587 ; Learning rate 1.0000 ; Source tokens/s 1822 ; Perplexity 264.24	
Epoch 2 ; Iteration 350/1587 ; Learning rate 1.0000 ; Source tokens/s 1805 ; Perplexity 260.15	
Epoch 2 ; Iteration 400/1587 ; Learning rate 1.0000 ; Source tokens/s 1797 ; Perplexity 255.19	
Epoch 2 ; Iteration 450/1587 ; Learning rate 1.0000 ; Source tokens/s 1801 ; Perplexity 250.83	
Epoch 2 ; Iteration 500/1587 ; Learning rate 1.0000 ; Source tokens/s 1803 ; Perplexity 246.35	
Epoch 2 ; Iteration 550/1587 ; Learning rate 1.0000 ; Source tokens/s 1820 ; Perplexity 243.61	
Epoch 2 ; Iteration 600/1587 ; Learning rate 1.0000 ; Source tokens/s 1820 ; Perplexity 240.35	
Epoch 2 ; Iteration 650/1587 ; Learning rate 1.0000 ; Source tokens/s 1826 ; Perplexity 237.36	
Epoch 2 ; Iteration 700/1587 ; Learning rate 1.0000 ; Source tokens/s 1821 ; Perplexity 233.45	
Epoch 2 ; Iteration 750/1587 ; Learning rate 1.0000 ; Source tokens/s 1821 ; Perplexity 230.06	
Epoch 2 ; Iteration 800/1587 ; Learning rate 1.0000 ; Source tokens/s 1826 ; Perplexity 226.20	
Epoch 2 ; Iteration 850/1587 ; Learning rate 1.0000 ; Source tokens/s 1825 ; Perplexity 222.75	
Epoch 2 ; Iteration 900/1587 ; Learning rate 1.0000 ; Source tokens/s 1836 ; Perplexity 220.01	
Epoch 2 ; Iteration 950/1587 ; Learning rate 1.0000 ; Source tokens/s 1838 ; Perplexity 216.68	
Epoch 2 ; Iteration 1000/1587 ; Learning rate 1.0000 ; Source tokens/s 1834 ; Perplexity 213.42	
Epoch 2 ; Iteration 1050/1587 ; Learning rate 1.0000 ; Source tokens/s 1835 ; Perplexity 210.49	
Epoch 2 ; Iteration 1100/1587 ; Learning rate 1.0000 ; Source tokens/s 1835 ; Perplexity 207.81	
Epoch 2 ; Iteration 1150/1587 ; Learning rate 1.0000 ; Source tokens/s 1833 ; Perplexity 204.49	
Epoch 2 ; Iteration 1200/1587 ; Learning rate 1.0000 ; Source tokens/s 1835 ; Perplexity 201.47	
Epoch 2 ; Iteration 1250/1587 ; Learning rate 1.0000 ; Source tokens/s 1834 ; Perplexity 198.52	
Epoch 2 ; Iteration 1300/1587 ; Learning rate 1.0000 ; Source tokens/s 1833 ; Perplexity 195.71	
Epoch 2 ; Iteration 1350/1587 ; Learning rate 1.0000 ; Source tokens/s 1838 ; Perplexity 192.95	
Epoch 2 ; Iteration 1400/1587 ; Learning rate 1.0000 ; Source tokens/s 1841 ; Perplexity 190.10	
Epoch 2 ; Iteration 1450/1587 ; Learning rate 1.0000 ; Source tokens/s 1839 ; Perplexity 187.43	
Epoch 2 ; Iteration 1500/1587 ; Learning rate 1.0000 ; Source tokens/s 1840 ; Perplexity 184.93	
Epoch 2 ; Iteration 1550/1587 ; Learning rate 1.0000 ; Source tokens/s 1837 ; Perplexity 182.65	
Validation perplexity: 94.961664704637	
Saving checkpoint to 'models/baseline_epoch2_94.96.t7'...	
	
Epoch 3 ; Iteration 50/1587 ; Learning rate 1.0000 ; Source tokens/s 1821 ; Perplexity 104.79	
Epoch 3 ; Iteration 100/1587 ; Learning rate 1.0000 ; Source tokens/s 1850 ; Perplexity 105.00	
Epoch 3 ; Iteration 150/1587 ; Learning rate 1.0000 ; Source tokens/s 1832 ; Perplexity 103.46	
Epoch 3 ; Iteration 200/1587 ; Learning rate 1.0000 ; Source tokens/s 1847 ; Perplexity 102.35	
Epoch 3 ; Iteration 250/1587 ; Learning rate 1.0000 ; Source tokens/s 1833 ; Perplexity 101.39	
Epoch 3 ; Iteration 300/1587 ; Learning rate 1.0000 ; Source tokens/s 1842 ; Perplexity 101.58	
Epoch 3 ; Iteration 350/1587 ; Learning rate 1.0000 ; Source tokens/s 1841 ; Perplexity 100.87	
Epoch 3 ; Iteration 400/1587 ; Learning rate 1.0000 ; Source tokens/s 1835 ; Perplexity 100.89	
Epoch 3 ; Iteration 450/1587 ; Learning rate 1.0000 ; Source tokens/s 1834 ; Perplexity 99.90	
Epoch 3 ; Iteration 500/1587 ; Learning rate 1.0000 ; Source tokens/s 1836 ; Perplexity 99.40	
Epoch 3 ; Iteration 550/1587 ; Learning rate 1.0000 ; Source tokens/s 1833 ; Perplexity 98.06	
Epoch 3 ; Iteration 600/1587 ; Learning rate 1.0000 ; Source tokens/s 1834 ; Perplexity 97.24	
Epoch 3 ; Iteration 650/1587 ; Learning rate 1.0000 ; Source tokens/s 1831 ; Perplexity 96.49	
Epoch 3 ; Iteration 700/1587 ; Learning rate 1.0000 ; Source tokens/s 1830 ; Perplexity 95.83	
Epoch 3 ; Iteration 750/1587 ; Learning rate 1.0000 ; Source tokens/s 1821 ; Perplexity 94.96	
Epoch 3 ; Iteration 800/1587 ; Learning rate 1.0000 ; Source tokens/s 1826 ; Perplexity 94.36	
Epoch 3 ; Iteration 850/1587 ; Learning rate 1.0000 ; Source tokens/s 1824 ; Perplexity 93.93	
Epoch 3 ; Iteration 900/1587 ; Learning rate 1.0000 ; Source tokens/s 1823 ; Perplexity 93.00	
Epoch 3 ; Iteration 950/1587 ; Learning rate 1.0000 ; Source tokens/s 1822 ; Perplexity 92.36	
Epoch 3 ; Iteration 1000/1587 ; Learning rate 1.0000 ; Source tokens/s 1825 ; Perplexity 91.76	
Epoch 3 ; Iteration 1050/1587 ; Learning rate 1.0000 ; Source tokens/s 1826 ; Perplexity 91.31	
Epoch 3 ; Iteration 1100/1587 ; Learning rate 1.0000 ; Source tokens/s 1826 ; Perplexity 90.59	
Epoch 3 ; Iteration 1150/1587 ; Learning rate 1.0000 ; Source tokens/s 1827 ; Perplexity 89.91	
Epoch 3 ; Iteration 1200/1587 ; Learning rate 1.0000 ; Source tokens/s 1827 ; Perplexity 89.37	
Epoch 3 ; Iteration 1250/1587 ; Learning rate 1.0000 ; Source tokens/s 1828 ; Perplexity 88.79	
Epoch 3 ; Iteration 1300/1587 ; Learning rate 1.0000 ; Source tokens/s 1835 ; Perplexity 88.23	
Epoch 3 ; Iteration 1350/1587 ; Learning rate 1.0000 ; Source tokens/s 1833 ; Perplexity 87.68	
Epoch 3 ; Iteration 1400/1587 ; Learning rate 1.0000 ; Source tokens/s 1831 ; Perplexity 87.09	
Epoch 3 ; Iteration 1450/1587 ; Learning rate 1.0000 ; Source tokens/s 1834 ; Perplexity 86.62	
Epoch 3 ; Iteration 1500/1587 ; Learning rate 1.0000 ; Source tokens/s 1837 ; Perplexity 86.11	
Epoch 3 ; Iteration 1550/1587 ; Learning rate 1.0000 ; Source tokens/s 1837 ; Perplexity 85.62	
Validation perplexity: 55.471133808158	
Saving checkpoint to 'models/baseline_epoch3_55.47.t7'...	
	
Epoch 4 ; Iteration 50/1587 ; Learning rate 1.0000 ; Source tokens/s 1928 ; Perplexity 60.31	
Epoch 4 ; Iteration 100/1587 ; Learning rate 1.0000 ; Source tokens/s 1864 ; Perplexity 59.71	
Epoch 4 ; Iteration 150/1587 ; Learning rate 1.0000 ; Source tokens/s 1869 ; Perplexity 60.24	
Epoch 4 ; Iteration 200/1587 ; Learning rate 1.0000 ; Source tokens/s 1863 ; Perplexity 59.44	
Epoch 4 ; Iteration 250/1587 ; Learning rate 1.0000 ; Source tokens/s 1863 ; Perplexity 59.33	
Epoch 4 ; Iteration 300/1587 ; Learning rate 1.0000 ; Source tokens/s 1858 ; Perplexity 58.78	
Epoch 4 ; Iteration 350/1587 ; Learning rate 1.0000 ; Source tokens/s 1784 ; Perplexity 58.58	
Epoch 4 ; Iteration 400/1587 ; Learning rate 1.0000 ; Source tokens/s 1564 ; Perplexity 58.83	
Epoch 4 ; Iteration 450/1587 ; Learning rate 1.0000 ; Source tokens/s 1415 ; Perplexity 58.76	
Epoch 4 ; Iteration 500/1587 ; Learning rate 1.0000 ; Source tokens/s 1334 ; Perplexity 58.35	
Epoch 4 ; Iteration 550/1587 ; Learning rate 1.0000 ; Source tokens/s 1264 ; Perplexity 58.25	
Epoch 4 ; Iteration 600/1587 ; Learning rate 1.0000 ; Source tokens/s 1212 ; Perplexity 58.03	
Epoch 4 ; Iteration 650/1587 ; Learning rate 1.0000 ; Source tokens/s 1169 ; Perplexity 57.93	
Epoch 4 ; Iteration 700/1587 ; Learning rate 1.0000 ; Source tokens/s 1133 ; Perplexity 57.86	
Epoch 4 ; Iteration 750/1587 ; Learning rate 1.0000 ; Source tokens/s 1108 ; Perplexity 57.66	
Epoch 4 ; Iteration 800/1587 ; Learning rate 1.0000 ; Source tokens/s 1084 ; Perplexity 57.55	
Epoch 4 ; Iteration 850/1587 ; Learning rate 1.0000 ; Source tokens/s 1064 ; Perplexity 57.44	
Epoch 4 ; Iteration 900/1587 ; Learning rate 1.0000 ; Source tokens/s 1047 ; Perplexity 57.39	
Epoch 4 ; Iteration 950/1587 ; Learning rate 1.0000 ; Source tokens/s 1034 ; Perplexity 57.12	
Epoch 4 ; Iteration 1000/1587 ; Learning rate 1.0000 ; Source tokens/s 1021 ; Perplexity 57.00	
Epoch 4 ; Iteration 1050/1587 ; Learning rate 1.0000 ; Source tokens/s 1012 ; Perplexity 56.93	
Epoch 4 ; Iteration 1100/1587 ; Learning rate 1.0000 ; Source tokens/s 1002 ; Perplexity 56.60	
Epoch 4 ; Iteration 1150/1587 ; Learning rate 1.0000 ; Source tokens/s 993 ; Perplexity 56.51	
Epoch 4 ; Iteration 1200/1587 ; Learning rate 1.0000 ; Source tokens/s 985 ; Perplexity 56.18	
Epoch 4 ; Iteration 1250/1587 ; Learning rate 1.0000 ; Source tokens/s 980 ; Perplexity 56.12	
Epoch 4 ; Iteration 1300/1587 ; Learning rate 1.0000 ; Source tokens/s 973 ; Perplexity 55.96	
Epoch 4 ; Iteration 1350/1587 ; Learning rate 1.0000 ; Source tokens/s 967 ; Perplexity 55.83	
Epoch 4 ; Iteration 1400/1587 ; Learning rate 1.0000 ; Source tokens/s 962 ; Perplexity 55.71	
Epoch 4 ; Iteration 1450/1587 ; Learning rate 1.0000 ; Source tokens/s 956 ; Perplexity 55.48	
Epoch 4 ; Iteration 1500/1587 ; Learning rate 1.0000 ; Source tokens/s 953 ; Perplexity 55.32	
Epoch 4 ; Iteration 1550/1587 ; Learning rate 1.0000 ; Source tokens/s 947 ; Perplexity 55.15	
Validation perplexity: 43.754144572934	
Saving checkpoint to 'models/baseline_epoch4_43.75.t7'...	
	
Epoch 5 ; Iteration 50/1587 ; Learning rate 1.0000 ; Source tokens/s 874 ; Perplexity 42.76	
Epoch 5 ; Iteration 100/1587 ; Learning rate 1.0000 ; Source tokens/s 850 ; Perplexity 41.09	
Epoch 5 ; Iteration 150/1587 ; Learning rate 1.0000 ; Source tokens/s 824 ; Perplexity 41.11	
Epoch 5 ; Iteration 200/1587 ; Learning rate 1.0000 ; Source tokens/s 831 ; Perplexity 40.90	
Epoch 5 ; Iteration 250/1587 ; Learning rate 1.0000 ; Source tokens/s 831 ; Perplexity 41.27	
Epoch 5 ; Iteration 300/1587 ; Learning rate 1.0000 ; Source tokens/s 836 ; Perplexity 41.87	
Epoch 5 ; Iteration 350/1587 ; Learning rate 1.0000 ; Source tokens/s 838 ; Perplexity 41.56	
Epoch 5 ; Iteration 400/1587 ; Learning rate 1.0000 ; Source tokens/s 836 ; Perplexity 41.56	
Epoch 5 ; Iteration 450/1587 ; Learning rate 1.0000 ; Source tokens/s 829 ; Perplexity 41.35	
Epoch 5 ; Iteration 500/1587 ; Learning rate 1.0000 ; Source tokens/s 832 ; Perplexity 41.30	
Epoch 5 ; Iteration 550/1587 ; Learning rate 1.0000 ; Source tokens/s 828 ; Perplexity 41.23	
Epoch 5 ; Iteration 600/1587 ; Learning rate 1.0000 ; Source tokens/s 830 ; Perplexity 41.34	
Epoch 5 ; Iteration 650/1587 ; Learning rate 1.0000 ; Source tokens/s 831 ; Perplexity 41.51	
Epoch 5 ; Iteration 700/1587 ; Learning rate 1.0000 ; Source tokens/s 832 ; Perplexity 41.60	
Epoch 5 ; Iteration 750/1587 ; Learning rate 1.0000 ; Source tokens/s 831 ; Perplexity 41.63	
Epoch 5 ; Iteration 800/1587 ; Learning rate 1.0000 ; Source tokens/s 832 ; Perplexity 41.76	
Epoch 5 ; Iteration 850/1587 ; Learning rate 1.0000 ; Source tokens/s 834 ; Perplexity 41.86	
Epoch 5 ; Iteration 900/1587 ; Learning rate 1.0000 ; Source tokens/s 834 ; Perplexity 41.73	
Epoch 5 ; Iteration 950/1587 ; Learning rate 1.0000 ; Source tokens/s 835 ; Perplexity 41.66	
Epoch 5 ; Iteration 1000/1587 ; Learning rate 1.0000 ; Source tokens/s 835 ; Perplexity 41.62	
Epoch 5 ; Iteration 1050/1587 ; Learning rate 1.0000 ; Source tokens/s 839 ; Perplexity 41.62	
Epoch 5 ; Iteration 1100/1587 ; Learning rate 1.0000 ; Source tokens/s 839 ; Perplexity 41.52	
Epoch 5 ; Iteration 1150/1587 ; Learning rate 1.0000 ; Source tokens/s 837 ; Perplexity 41.48	
Epoch 5 ; Iteration 1200/1587 ; Learning rate 1.0000 ; Source tokens/s 835 ; Perplexity 41.38	
Epoch 5 ; Iteration 1250/1587 ; Learning rate 1.0000 ; Source tokens/s 837 ; Perplexity 41.30	
Epoch 5 ; Iteration 1300/1587 ; Learning rate 1.0000 ; Source tokens/s 837 ; Perplexity 41.30	
Epoch 5 ; Iteration 1350/1587 ; Learning rate 1.0000 ; Source tokens/s 838 ; Perplexity 41.20	
Epoch 5 ; Iteration 1400/1587 ; Learning rate 1.0000 ; Source tokens/s 839 ; Perplexity 41.12	
Epoch 5 ; Iteration 1450/1587 ; Learning rate 1.0000 ; Source tokens/s 842 ; Perplexity 41.14	
Epoch 5 ; Iteration 1500/1587 ; Learning rate 1.0000 ; Source tokens/s 843 ; Perplexity 41.12	
Epoch 5 ; Iteration 1550/1587 ; Learning rate 1.0000 ; Source tokens/s 840 ; Perplexity 41.07	
Validation perplexity: 35.248375765707	
Saving checkpoint to 'models/baseline_epoch5_35.25.t7'...	
	
Epoch 6 ; Iteration 50/1587 ; Learning rate 1.0000 ; Source tokens/s 811 ; Perplexity 32.54	
Epoch 6 ; Iteration 100/1587 ; Learning rate 1.0000 ; Source tokens/s 807 ; Perplexity 32.51	
Epoch 6 ; Iteration 150/1587 ; Learning rate 1.0000 ; Source tokens/s 818 ; Perplexity 32.59	
Epoch 6 ; Iteration 200/1587 ; Learning rate 1.0000 ; Source tokens/s 801 ; Perplexity 32.15	
Epoch 6 ; Iteration 250/1587 ; Learning rate 1.0000 ; Source tokens/s 812 ; Perplexity 32.23	
Epoch 6 ; Iteration 300/1587 ; Learning rate 1.0000 ; Source tokens/s 818 ; Perplexity 32.35	
Epoch 6 ; Iteration 350/1587 ; Learning rate 1.0000 ; Source tokens/s 830 ; Perplexity 32.59	
Epoch 6 ; Iteration 400/1587 ; Learning rate 1.0000 ; Source tokens/s 831 ; Perplexity 32.73	
Epoch 6 ; Iteration 450/1587 ; Learning rate 1.0000 ; Source tokens/s 831 ; Perplexity 32.69	
Epoch 6 ; Iteration 500/1587 ; Learning rate 1.0000 ; Source tokens/s 829 ; Perplexity 32.74	
Epoch 6 ; Iteration 550/1587 ; Learning rate 1.0000 ; Source tokens/s 829 ; Perplexity 32.93	
Epoch 6 ; Iteration 600/1587 ; Learning rate 1.0000 ; Source tokens/s 829 ; Perplexity 33.09	
Epoch 6 ; Iteration 650/1587 ; Learning rate 1.0000 ; Source tokens/s 830 ; Perplexity 32.98	
Epoch 6 ; Iteration 700/1587 ; Learning rate 1.0000 ; Source tokens/s 829 ; Perplexity 32.94	
Epoch 6 ; Iteration 750/1587 ; Learning rate 1.0000 ; Source tokens/s 831 ; Perplexity 32.84	
Epoch 6 ; Iteration 800/1587 ; Learning rate 1.0000 ; Source tokens/s 831 ; Perplexity 32.74	
Epoch 6 ; Iteration 850/1587 ; Learning rate 1.0000 ; Source tokens/s 831 ; Perplexity 32.69	
Epoch 6 ; Iteration 900/1587 ; Learning rate 1.0000 ; Source tokens/s 831 ; Perplexity 32.63	
Epoch 6 ; Iteration 950/1587 ; Learning rate 1.0000 ; Source tokens/s 832 ; Perplexity 32.73	
Epoch 6 ; Iteration 1000/1587 ; Learning rate 1.0000 ; Source tokens/s 832 ; Perplexity 32.67	
Epoch 6 ; Iteration 1050/1587 ; Learning rate 1.0000 ; Source tokens/s 832 ; Perplexity 32.71	
Epoch 6 ; Iteration 1100/1587 ; Learning rate 1.0000 ; Source tokens/s 831 ; Perplexity 32.67	
Epoch 6 ; Iteration 1150/1587 ; Learning rate 1.0000 ; Source tokens/s 834 ; Perplexity 32.79	
Epoch 6 ; Iteration 1200/1587 ; Learning rate 1.0000 ; Source tokens/s 834 ; Perplexity 32.82	
Epoch 6 ; Iteration 1250/1587 ; Learning rate 1.0000 ; Source tokens/s 833 ; Perplexity 32.81	
Epoch 6 ; Iteration 1300/1587 ; Learning rate 1.0000 ; Source tokens/s 832 ; Perplexity 32.84	
Epoch 6 ; Iteration 1350/1587 ; Learning rate 1.0000 ; Source tokens/s 832 ; Perplexity 32.84	
Epoch 6 ; Iteration 1400/1587 ; Learning rate 1.0000 ; Source tokens/s 833 ; Perplexity 32.88	
Epoch 6 ; Iteration 1450/1587 ; Learning rate 1.0000 ; Source tokens/s 833 ; Perplexity 32.91	
Epoch 6 ; Iteration 1500/1587 ; Learning rate 1.0000 ; Source tokens/s 834 ; Perplexity 32.91	
Epoch 6 ; Iteration 1550/1587 ; Learning rate 1.0000 ; Source tokens/s 835 ; Perplexity 32.89	
Validation perplexity: 31.98985843374	
Saving checkpoint to 'models/baseline_epoch6_31.99.t7'...	
	
Epoch 7 ; Iteration 50/1587 ; Learning rate 1.0000 ; Source tokens/s 846 ; Perplexity 27.00	
Epoch 7 ; Iteration 100/1587 ; Learning rate 1.0000 ; Source tokens/s 864 ; Perplexity 26.83	
Epoch 7 ; Iteration 150/1587 ; Learning rate 1.0000 ; Source tokens/s 847 ; Perplexity 26.66	
Epoch 7 ; Iteration 200/1587 ; Learning rate 1.0000 ; Source tokens/s 839 ; Perplexity 26.58	
Epoch 7 ; Iteration 250/1587 ; Learning rate 1.0000 ; Source tokens/s 846 ; Perplexity 26.90	
Epoch 7 ; Iteration 300/1587 ; Learning rate 1.0000 ; Source tokens/s 841 ; Perplexity 26.73	
Epoch 7 ; Iteration 350/1587 ; Learning rate 1.0000 ; Source tokens/s 844 ; Perplexity 26.69	
Epoch 7 ; Iteration 400/1587 ; Learning rate 1.0000 ; Source tokens/s 842 ; Perplexity 26.64	
Epoch 7 ; Iteration 450/1587 ; Learning rate 1.0000 ; Source tokens/s 843 ; Perplexity 26.70	
Epoch 7 ; Iteration 500/1587 ; Learning rate 1.0000 ; Source tokens/s 843 ; Perplexity 26.71	
Epoch 7 ; Iteration 550/1587 ; Learning rate 1.0000 ; Source tokens/s 843 ; Perplexity 26.69	
Epoch 7 ; Iteration 600/1587 ; Learning rate 1.0000 ; Source tokens/s 838 ; Perplexity 26.82	
Epoch 7 ; Iteration 650/1587 ; Learning rate 1.0000 ; Source tokens/s 838 ; Perplexity 27.01	
Epoch 7 ; Iteration 700/1587 ; Learning rate 1.0000 ; Source tokens/s 834 ; Perplexity 26.99	
Epoch 7 ; Iteration 750/1587 ; Learning rate 1.0000 ; Source tokens/s 836 ; Perplexity 27.11	
Epoch 7 ; Iteration 800/1587 ; Learning rate 1.0000 ; Source tokens/s 834 ; Perplexity 27.15	
Epoch 7 ; Iteration 850/1587 ; Learning rate 1.0000 ; Source tokens/s 836 ; Perplexity 27.09	
Epoch 7 ; Iteration 900/1587 ; Learning rate 1.0000 ; Source tokens/s 840 ; Perplexity 27.18	
Epoch 7 ; Iteration 950/1587 ; Learning rate 1.0000 ; Source tokens/s 838 ; Perplexity 27.17	
Epoch 7 ; Iteration 1000/1587 ; Learning rate 1.0000 ; Source tokens/s 836 ; Perplexity 27.20	
Epoch 7 ; Iteration 1050/1587 ; Learning rate 1.0000 ; Source tokens/s 836 ; Perplexity 27.30	
Epoch 7 ; Iteration 1100/1587 ; Learning rate 1.0000 ; Source tokens/s 839 ; Perplexity 27.31	
Epoch 7 ; Iteration 1150/1587 ; Learning rate 1.0000 ; Source tokens/s 839 ; Perplexity 27.35	
Epoch 7 ; Iteration 1200/1587 ; Learning rate 1.0000 ; Source tokens/s 839 ; Perplexity 27.39	
Epoch 7 ; Iteration 1250/1587 ; Learning rate 1.0000 ; Source tokens/s 840 ; Perplexity 27.40	
Epoch 7 ; Iteration 1300/1587 ; Learning rate 1.0000 ; Source tokens/s 841 ; Perplexity 27.43	
Epoch 7 ; Iteration 1350/1587 ; Learning rate 1.0000 ; Source tokens/s 840 ; Perplexity 27.42	
Epoch 7 ; Iteration 1400/1587 ; Learning rate 1.0000 ; Source tokens/s 841 ; Perplexity 27.45	
Epoch 7 ; Iteration 1450/1587 ; Learning rate 1.0000 ; Source tokens/s 840 ; Perplexity 27.48	
Epoch 7 ; Iteration 1500/1587 ; Learning rate 1.0000 ; Source tokens/s 840 ; Perplexity 27.50	
Epoch 7 ; Iteration 1550/1587 ; Learning rate 1.0000 ; Source tokens/s 840 ; Perplexity 27.50	
Validation perplexity: 28.895515910504	
Saving checkpoint to 'models/baseline_epoch7_28.90.t7'...	
	
Epoch 8 ; Iteration 50/1587 ; Learning rate 1.0000 ; Source tokens/s 823 ; Perplexity 22.72	
Epoch 8 ; Iteration 100/1587 ; Learning rate 1.0000 ; Source tokens/s 839 ; Perplexity 22.44	
Epoch 8 ; Iteration 150/1587 ; Learning rate 1.0000 ; Source tokens/s 863 ; Perplexity 22.41	
Epoch 8 ; Iteration 200/1587 ; Learning rate 1.0000 ; Source tokens/s 855 ; Perplexity 22.78	
Epoch 8 ; Iteration 250/1587 ; Learning rate 1.0000 ; Source tokens/s 849 ; Perplexity 22.36	
Epoch 8 ; Iteration 300/1587 ; Learning rate 1.0000 ; Source tokens/s 853 ; Perplexity 22.67	
Epoch 8 ; Iteration 350/1587 ; Learning rate 1.0000 ; Source tokens/s 851 ; Perplexity 22.65	
Epoch 8 ; Iteration 400/1587 ; Learning rate 1.0000 ; Source tokens/s 849 ; Perplexity 22.61	
Epoch 8 ; Iteration 450/1587 ; Learning rate 1.0000 ; Source tokens/s 847 ; Perplexity 22.70	
Epoch 8 ; Iteration 500/1587 ; Learning rate 1.0000 ; Source tokens/s 846 ; Perplexity 22.68	
Epoch 8 ; Iteration 550/1587 ; Learning rate 1.0000 ; Source tokens/s 843 ; Perplexity 22.76	
Epoch 8 ; Iteration 600/1587 ; Learning rate 1.0000 ; Source tokens/s 841 ; Perplexity 22.86	
Epoch 8 ; Iteration 650/1587 ; Learning rate 1.0000 ; Source tokens/s 844 ; Perplexity 23.01	
Epoch 8 ; Iteration 700/1587 ; Learning rate 1.0000 ; Source tokens/s 844 ; Perplexity 23.06	
Epoch 8 ; Iteration 750/1587 ; Learning rate 1.0000 ; Source tokens/s 844 ; Perplexity 23.13	
Epoch 8 ; Iteration 800/1587 ; Learning rate 1.0000 ; Source tokens/s 845 ; Perplexity 23.19	
Epoch 8 ; Iteration 850/1587 ; Learning rate 1.0000 ; Source tokens/s 843 ; Perplexity 23.17	
Epoch 8 ; Iteration 900/1587 ; Learning rate 1.0000 ; Source tokens/s 838 ; Perplexity 23.15	
Epoch 8 ; Iteration 950/1587 ; Learning rate 1.0000 ; Source tokens/s 840 ; Perplexity 23.27	
Epoch 8 ; Iteration 1000/1587 ; Learning rate 1.0000 ; Source tokens/s 838 ; Perplexity 23.28	
Epoch 8 ; Iteration 1050/1587 ; Learning rate 1.0000 ; Source tokens/s 838 ; Perplexity 23.37	
Epoch 8 ; Iteration 1100/1587 ; Learning rate 1.0000 ; Source tokens/s 838 ; Perplexity 23.48	
Epoch 8 ; Iteration 1150/1587 ; Learning rate 1.0000 ; Source tokens/s 836 ; Perplexity 23.49	
Epoch 8 ; Iteration 1200/1587 ; Learning rate 1.0000 ; Source tokens/s 835 ; Perplexity 23.47	
Epoch 8 ; Iteration 1250/1587 ; Learning rate 1.0000 ; Source tokens/s 836 ; Perplexity 23.48	
Epoch 8 ; Iteration 1300/1587 ; Learning rate 1.0000 ; Source tokens/s 835 ; Perplexity 23.48	
Epoch 8 ; Iteration 1350/1587 ; Learning rate 1.0000 ; Source tokens/s 836 ; Perplexity 23.55	
Epoch 8 ; Iteration 1400/1587 ; Learning rate 1.0000 ; Source tokens/s 837 ; Perplexity 23.55	
Epoch 8 ; Iteration 1450/1587 ; Learning rate 1.0000 ; Source tokens/s 835 ; Perplexity 23.56	
Epoch 8 ; Iteration 1500/1587 ; Learning rate 1.0000 ; Source tokens/s 837 ; Perplexity 23.62	
Epoch 8 ; Iteration 1550/1587 ; Learning rate 1.0000 ; Source tokens/s 836 ; Perplexity 23.60	
Validation perplexity: 27.403117648073	
Saving checkpoint to 'models/baseline_epoch8_27.40.t7'...	
	
Epoch 9 ; Iteration 50/1587 ; Learning rate 1.0000 ; Source tokens/s 856 ; Perplexity 19.06	
Epoch 9 ; Iteration 100/1587 ; Learning rate 1.0000 ; Source tokens/s 832 ; Perplexity 19.52	
Epoch 9 ; Iteration 150/1587 ; Learning rate 1.0000 ; Source tokens/s 822 ; Perplexity 19.22	
Epoch 9 ; Iteration 200/1587 ; Learning rate 1.0000 ; Source tokens/s 814 ; Perplexity 19.31	
Epoch 9 ; Iteration 250/1587 ; Learning rate 1.0000 ; Source tokens/s 837 ; Perplexity 19.79	
Epoch 9 ; Iteration 300/1587 ; Learning rate 1.0000 ; Source tokens/s 850 ; Perplexity 19.65	
Epoch 9 ; Iteration 350/1587 ; Learning rate 1.0000 ; Source tokens/s 842 ; Perplexity 19.66	
Epoch 9 ; Iteration 400/1587 ; Learning rate 1.0000 ; Source tokens/s 836 ; Perplexity 19.62	
Epoch 9 ; Iteration 450/1587 ; Learning rate 1.0000 ; Source tokens/s 837 ; Perplexity 19.76	
Epoch 9 ; Iteration 500/1587 ; Learning rate 1.0000 ; Source tokens/s 844 ; Perplexity 19.93	
Epoch 9 ; Iteration 550/1587 ; Learning rate 1.0000 ; Source tokens/s 841 ; Perplexity 19.95	
Epoch 9 ; Iteration 600/1587 ; Learning rate 1.0000 ; Source tokens/s 839 ; Perplexity 19.93	
Epoch 9 ; Iteration 650/1587 ; Learning rate 1.0000 ; Source tokens/s 837 ; Perplexity 19.93	
Epoch 9 ; Iteration 700/1587 ; Learning rate 1.0000 ; Source tokens/s 840 ; Perplexity 20.04	
Epoch 9 ; Iteration 750/1587 ; Learning rate 1.0000 ; Source tokens/s 839 ; Perplexity 20.09	
Epoch 9 ; Iteration 800/1587 ; Learning rate 1.0000 ; Source tokens/s 843 ; Perplexity 20.21	
Epoch 9 ; Iteration 850/1587 ; Learning rate 1.0000 ; Source tokens/s 838 ; Perplexity 20.21	
Epoch 9 ; Iteration 900/1587 ; Learning rate 1.0000 ; Source tokens/s 841 ; Perplexity 20.33	
Epoch 9 ; Iteration 950/1587 ; Learning rate 1.0000 ; Source tokens/s 843 ; Perplexity 20.33	
Epoch 9 ; Iteration 1000/1587 ; Learning rate 1.0000 ; Source tokens/s 842 ; Perplexity 20.38	
Epoch 9 ; Iteration 1050/1587 ; Learning rate 1.0000 ; Source tokens/s 841 ; Perplexity 20.37	
Epoch 9 ; Iteration 1100/1587 ; Learning rate 1.0000 ; Source tokens/s 839 ; Perplexity 20.43	
Epoch 9 ; Iteration 1150/1587 ; Learning rate 1.0000 ; Source tokens/s 840 ; Perplexity 20.47	
Epoch 9 ; Iteration 1200/1587 ; Learning rate 1.0000 ; Source tokens/s 839 ; Perplexity 20.51	
Epoch 9 ; Iteration 1250/1587 ; Learning rate 1.0000 ; Source tokens/s 837 ; Perplexity 20.48	
Epoch 9 ; Iteration 1300/1587 ; Learning rate 1.0000 ; Source tokens/s 837 ; Perplexity 20.48	
Epoch 9 ; Iteration 1350/1587 ; Learning rate 1.0000 ; Source tokens/s 837 ; Perplexity 20.52	
Epoch 9 ; Iteration 1400/1587 ; Learning rate 1.0000 ; Source tokens/s 837 ; Perplexity 20.54	
Epoch 9 ; Iteration 1450/1587 ; Learning rate 1.0000 ; Source tokens/s 837 ; Perplexity 20.56	
Epoch 9 ; Iteration 1500/1587 ; Learning rate 1.0000 ; Source tokens/s 838 ; Perplexity 20.59	
Epoch 9 ; Iteration 1550/1587 ; Learning rate 1.0000 ; Source tokens/s 839 ; Perplexity 20.64	
Validation perplexity: 25.825762106436	
Saving checkpoint to 'models/baseline_epoch9_25.83.t7'...	
	
Epoch 10 ; Iteration 50/1587 ; Learning rate 0.5000 ; Source tokens/s 818 ; Perplexity 15.92	
Epoch 10 ; Iteration 100/1587 ; Learning rate 0.5000 ; Source tokens/s 822 ; Perplexity 15.73	
Epoch 10 ; Iteration 150/1587 ; Learning rate 0.5000 ; Source tokens/s 817 ; Perplexity 15.87	
Epoch 10 ; Iteration 200/1587 ; Learning rate 0.5000 ; Source tokens/s 823 ; Perplexity 15.53	
Epoch 10 ; Iteration 250/1587 ; Learning rate 0.5000 ; Source tokens/s 827 ; Perplexity 15.60	
Epoch 10 ; Iteration 300/1587 ; Learning rate 0.5000 ; Source tokens/s 826 ; Perplexity 15.51	
Epoch 10 ; Iteration 350/1587 ; Learning rate 0.5000 ; Source tokens/s 829 ; Perplexity 15.45	
Epoch 10 ; Iteration 400/1587 ; Learning rate 0.5000 ; Source tokens/s 820 ; Perplexity 15.37	
Epoch 10 ; Iteration 450/1587 ; Learning rate 0.5000 ; Source tokens/s 819 ; Perplexity 15.29	
Epoch 10 ; Iteration 500/1587 ; Learning rate 0.5000 ; Source tokens/s 817 ; Perplexity 15.23	
Epoch 10 ; Iteration 550/1587 ; Learning rate 0.5000 ; Source tokens/s 820 ; Perplexity 15.21	
Epoch 10 ; Iteration 600/1587 ; Learning rate 0.5000 ; Source tokens/s 822 ; Perplexity 15.28	
Epoch 10 ; Iteration 650/1587 ; Learning rate 0.5000 ; Source tokens/s 822 ; Perplexity 15.29	
Epoch 10 ; Iteration 700/1587 ; Learning rate 0.5000 ; Source tokens/s 824 ; Perplexity 15.30	
Epoch 10 ; Iteration 750/1587 ; Learning rate 0.5000 ; Source tokens/s 826 ; Perplexity 15.26	
Epoch 10 ; Iteration 800/1587 ; Learning rate 0.5000 ; Source tokens/s 827 ; Perplexity 15.31	
Epoch 10 ; Iteration 850/1587 ; Learning rate 0.5000 ; Source tokens/s 827 ; Perplexity 15.32	
Epoch 10 ; Iteration 900/1587 ; Learning rate 0.5000 ; Source tokens/s 826 ; Perplexity 15.33	
Epoch 10 ; Iteration 950/1587 ; Learning rate 0.5000 ; Source tokens/s 826 ; Perplexity 15.35	
Epoch 10 ; Iteration 1000/1587 ; Learning rate 0.5000 ; Source tokens/s 828 ; Perplexity 15.41	
Epoch 10 ; Iteration 1050/1587 ; Learning rate 0.5000 ; Source tokens/s 832 ; Perplexity 15.45	
Epoch 10 ; Iteration 1100/1587 ; Learning rate 0.5000 ; Source tokens/s 832 ; Perplexity 15.48	
Epoch 10 ; Iteration 1150/1587 ; Learning rate 0.5000 ; Source tokens/s 833 ; Perplexity 15.43	
Epoch 10 ; Iteration 1200/1587 ; Learning rate 0.5000 ; Source tokens/s 832 ; Perplexity 15.42	
Epoch 10 ; Iteration 1250/1587 ; Learning rate 0.5000 ; Source tokens/s 832 ; Perplexity 15.41	
Epoch 10 ; Iteration 1300/1587 ; Learning rate 0.5000 ; Source tokens/s 833 ; Perplexity 15.44	
Epoch 10 ; Iteration 1350/1587 ; Learning rate 0.5000 ; Source tokens/s 836 ; Perplexity 15.46	
Epoch 10 ; Iteration 1400/1587 ; Learning rate 0.5000 ; Source tokens/s 839 ; Perplexity 15.49	
Epoch 10 ; Iteration 1450/1587 ; Learning rate 0.5000 ; Source tokens/s 840 ; Perplexity 15.48	
Epoch 10 ; Iteration 1500/1587 ; Learning rate 0.5000 ; Source tokens/s 841 ; Perplexity 15.49	
Epoch 10 ; Iteration 1550/1587 ; Learning rate 0.5000 ; Source tokens/s 840 ; Perplexity 15.50	
Validation perplexity: 23.989800570611	
Saving checkpoint to 'models/baseline_epoch10_23.99.t7'...	
	
Epoch 11 ; Iteration 50/1587 ; Learning rate 0.2500 ; Source tokens/s 780 ; Perplexity 12.87	
Epoch 11 ; Iteration 100/1587 ; Learning rate 0.2500 ; Source tokens/s 831 ; Perplexity 13.25	
Epoch 11 ; Iteration 150/1587 ; Learning rate 0.2500 ; Source tokens/s 839 ; Perplexity 13.02	
Epoch 11 ; Iteration 200/1587 ; Learning rate 0.2500 ; Source tokens/s 844 ; Perplexity 12.91	
Epoch 11 ; Iteration 250/1587 ; Learning rate 0.2500 ; Source tokens/s 847 ; Perplexity 13.01	
Epoch 11 ; Iteration 300/1587 ; Learning rate 0.2500 ; Source tokens/s 841 ; Perplexity 12.95	
Epoch 11 ; Iteration 350/1587 ; Learning rate 0.2500 ; Source tokens/s 839 ; Perplexity 12.90	
Epoch 11 ; Iteration 400/1587 ; Learning rate 0.2500 ; Source tokens/s 836 ; Perplexity 12.89	
Epoch 11 ; Iteration 450/1587 ; Learning rate 0.2500 ; Source tokens/s 833 ; Perplexity 12.87	
Epoch 11 ; Iteration 500/1587 ; Learning rate 0.2500 ; Source tokens/s 835 ; Perplexity 12.79	
Epoch 11 ; Iteration 550/1587 ; Learning rate 0.2500 ; Source tokens/s 834 ; Perplexity 12.77	
Epoch 11 ; Iteration 600/1587 ; Learning rate 0.2500 ; Source tokens/s 839 ; Perplexity 12.80	
Epoch 11 ; Iteration 650/1587 ; Learning rate 0.2500 ; Source tokens/s 843 ; Perplexity 12.86	
Epoch 11 ; Iteration 700/1587 ; Learning rate 0.2500 ; Source tokens/s 842 ; Perplexity 12.89	
Epoch 11 ; Iteration 750/1587 ; Learning rate 0.2500 ; Source tokens/s 842 ; Perplexity 12.87	
Epoch 11 ; Iteration 800/1587 ; Learning rate 0.2500 ; Source tokens/s 839 ; Perplexity 12.86	
Epoch 11 ; Iteration 850/1587 ; Learning rate 0.2500 ; Source tokens/s 841 ; Perplexity 12.89	
Epoch 11 ; Iteration 900/1587 ; Learning rate 0.2500 ; Source tokens/s 838 ; Perplexity 12.86	
Epoch 11 ; Iteration 950/1587 ; Learning rate 0.2500 ; Source tokens/s 837 ; Perplexity 12.83	
Epoch 11 ; Iteration 1000/1587 ; Learning rate 0.2500 ; Source tokens/s 839 ; Perplexity 12.85	
Epoch 11 ; Iteration 1050/1587 ; Learning rate 0.2500 ; Source tokens/s 837 ; Perplexity 12.80	
Epoch 11 ; Iteration 1100/1587 ; Learning rate 0.2500 ; Source tokens/s 837 ; Perplexity 12.80	
Epoch 11 ; Iteration 1150/1587 ; Learning rate 0.2500 ; Source tokens/s 838 ; Perplexity 12.85	
Epoch 11 ; Iteration 1200/1587 ; Learning rate 0.2500 ; Source tokens/s 838 ; Perplexity 12.84	
Epoch 11 ; Iteration 1250/1587 ; Learning rate 0.2500 ; Source tokens/s 837 ; Perplexity 12.83	
Epoch 11 ; Iteration 1300/1587 ; Learning rate 0.2500 ; Source tokens/s 838 ; Perplexity 12.84	
Epoch 11 ; Iteration 1350/1587 ; Learning rate 0.2500 ; Source tokens/s 837 ; Perplexity 12.86	
Epoch 11 ; Iteration 1400/1587 ; Learning rate 0.2500 ; Source tokens/s 837 ; Perplexity 12.86	
Epoch 11 ; Iteration 1450/1587 ; Learning rate 0.2500 ; Source tokens/s 837 ; Perplexity 12.87	
Epoch 11 ; Iteration 1500/1587 ; Learning rate 0.2500 ; Source tokens/s 837 ; Perplexity 12.86	
Epoch 11 ; Iteration 1550/1587 ; Learning rate 0.2500 ; Source tokens/s 835 ; Perplexity 12.83	
Validation perplexity: 22.960085550097	
Saving checkpoint to 'models/baseline_epoch11_22.96.t7'...	
	
Epoch 12 ; Iteration 50/1587 ; Learning rate 0.1250 ; Source tokens/s 863 ; Perplexity 12.03	
Epoch 12 ; Iteration 100/1587 ; Learning rate 0.1250 ; Source tokens/s 835 ; Perplexity 11.52	
Epoch 12 ; Iteration 150/1587 ; Learning rate 0.1250 ; Source tokens/s 847 ; Perplexity 11.65	
Epoch 12 ; Iteration 200/1587 ; Learning rate 0.1250 ; Source tokens/s 845 ; Perplexity 11.59	
Epoch 12 ; Iteration 250/1587 ; Learning rate 0.1250 ; Source tokens/s 853 ; Perplexity 11.71	
Epoch 12 ; Iteration 300/1587 ; Learning rate 0.1250 ; Source tokens/s 852 ; Perplexity 11.88	
Epoch 12 ; Iteration 350/1587 ; Learning rate 0.1250 ; Source tokens/s 849 ; Perplexity 11.82	
Epoch 12 ; Iteration 400/1587 ; Learning rate 0.1250 ; Source tokens/s 841 ; Perplexity 11.70	
Epoch 12 ; Iteration 450/1587 ; Learning rate 0.1250 ; Source tokens/s 838 ; Perplexity 11.64	
Epoch 12 ; Iteration 500/1587 ; Learning rate 0.1250 ; Source tokens/s 839 ; Perplexity 11.69	
Epoch 12 ; Iteration 550/1587 ; Learning rate 0.1250 ; Source tokens/s 835 ; Perplexity 11.57	
Epoch 12 ; Iteration 600/1587 ; Learning rate 0.1250 ; Source tokens/s 836 ; Perplexity 11.58	
Epoch 12 ; Iteration 650/1587 ; Learning rate 0.1250 ; Source tokens/s 835 ; Perplexity 11.60	
Epoch 12 ; Iteration 700/1587 ; Learning rate 0.1250 ; Source tokens/s 835 ; Perplexity 11.57	
Epoch 12 ; Iteration 750/1587 ; Learning rate 0.1250 ; Source tokens/s 836 ; Perplexity 11.62	
Epoch 12 ; Iteration 800/1587 ; Learning rate 0.1250 ; Source tokens/s 841 ; Perplexity 11.63	
Epoch 12 ; Iteration 850/1587 ; Learning rate 0.1250 ; Source tokens/s 843 ; Perplexity 11.67	
Epoch 12 ; Iteration 900/1587 ; Learning rate 0.1250 ; Source tokens/s 841 ; Perplexity 11.65	
Epoch 12 ; Iteration 950/1587 ; Learning rate 0.1250 ; Source tokens/s 845 ; Perplexity 11.67	
Epoch 12 ; Iteration 1000/1587 ; Learning rate 0.1250 ; Source tokens/s 844 ; Perplexity 11.68	
Epoch 12 ; Iteration 1050/1587 ; Learning rate 0.1250 ; Source tokens/s 843 ; Perplexity 11.66	
Epoch 12 ; Iteration 1100/1587 ; Learning rate 0.1250 ; Source tokens/s 842 ; Perplexity 11.65	
Epoch 12 ; Iteration 1150/1587 ; Learning rate 0.1250 ; Source tokens/s 842 ; Perplexity 11.65	
Epoch 12 ; Iteration 1200/1587 ; Learning rate 0.1250 ; Source tokens/s 841 ; Perplexity 11.64	
Epoch 12 ; Iteration 1250/1587 ; Learning rate 0.1250 ; Source tokens/s 843 ; Perplexity 11.66	
Epoch 12 ; Iteration 1300/1587 ; Learning rate 0.1250 ; Source tokens/s 842 ; Perplexity 11.64	
Epoch 12 ; Iteration 1350/1587 ; Learning rate 0.1250 ; Source tokens/s 841 ; Perplexity 11.61	
Epoch 12 ; Iteration 1400/1587 ; Learning rate 0.1250 ; Source tokens/s 841 ; Perplexity 11.61	
Epoch 12 ; Iteration 1450/1587 ; Learning rate 0.1250 ; Source tokens/s 842 ; Perplexity 11.63	
Epoch 12 ; Iteration 1500/1587 ; Learning rate 0.1250 ; Source tokens/s 841 ; Perplexity 11.61	
Epoch 12 ; Iteration 1550/1587 ; Learning rate 0.1250 ; Source tokens/s 839 ; Perplexity 11.58	
Validation perplexity: 22.833532364382	
Saving checkpoint to 'models/baseline_epoch12_22.83.t7'...	
	
Epoch 13 ; Iteration 50/1587 ; Learning rate 0.0625 ; Source tokens/s 815 ; Perplexity 10.74	
Epoch 13 ; Iteration 100/1587 ; Learning rate 0.0625 ; Source tokens/s 824 ; Perplexity 11.03	
Epoch 13 ; Iteration 150/1587 ; Learning rate 0.0625 ; Source tokens/s 828 ; Perplexity 11.10	
Epoch 13 ; Iteration 200/1587 ; Learning rate 0.0625 ; Source tokens/s 821 ; Perplexity 10.92	
Epoch 13 ; Iteration 250/1587 ; Learning rate 0.0625 ; Source tokens/s 828 ; Perplexity 10.89	
Epoch 13 ; Iteration 300/1587 ; Learning rate 0.0625 ; Source tokens/s 832 ; Perplexity 11.01	
Epoch 13 ; Iteration 350/1587 ; Learning rate 0.0625 ; Source tokens/s 830 ; Perplexity 10.93	
Epoch 13 ; Iteration 400/1587 ; Learning rate 0.0625 ; Source tokens/s 831 ; Perplexity 10.95	
Epoch 13 ; Iteration 450/1587 ; Learning rate 0.0625 ; Source tokens/s 829 ; Perplexity 10.93	
Epoch 13 ; Iteration 500/1587 ; Learning rate 0.0625 ; Source tokens/s 829 ; Perplexity 10.94	
Epoch 13 ; Iteration 550/1587 ; Learning rate 0.0625 ; Source tokens/s 828 ; Perplexity 10.90	
Epoch 13 ; Iteration 600/1587 ; Learning rate 0.0625 ; Source tokens/s 829 ; Perplexity 10.89	
Epoch 13 ; Iteration 650/1587 ; Learning rate 0.0625 ; Source tokens/s 843 ; Perplexity 10.89	
Epoch 13 ; Iteration 700/1587 ; Learning rate 0.0625 ; Source tokens/s 877 ; Perplexity 10.84	
Epoch 13 ; Iteration 750/1587 ; Learning rate 0.0625 ; Source tokens/s 906 ; Perplexity 10.85	
Epoch 13 ; Iteration 800/1587 ; Learning rate 0.0625 ; Source tokens/s 939 ; Perplexity 10.89	
Epoch 13 ; Iteration 850/1587 ; Learning rate 0.0625 ; Source tokens/s 930 ; Perplexity 10.88	
Epoch 13 ; Iteration 900/1587 ; Learning rate 0.0625 ; Source tokens/s 921 ; Perplexity 10.85	
Epoch 13 ; Iteration 950/1587 ; Learning rate 0.0625 ; Source tokens/s 916 ; Perplexity 10.82	
Epoch 13 ; Iteration 1000/1587 ; Learning rate 0.0625 ; Source tokens/s 910 ; Perplexity 10.85	
Epoch 13 ; Iteration 1050/1587 ; Learning rate 0.0625 ; Source tokens/s 904 ; Perplexity 10.82	
Epoch 13 ; Iteration 1100/1587 ; Learning rate 0.0625 ; Source tokens/s 899 ; Perplexity 10.81	
Epoch 13 ; Iteration 1150/1587 ; Learning rate 0.0625 ; Source tokens/s 897 ; Perplexity 10.83	
Epoch 13 ; Iteration 1200/1587 ; Learning rate 0.0625 ; Source tokens/s 896 ; Perplexity 10.89	
Epoch 13 ; Iteration 1250/1587 ; Learning rate 0.0625 ; Source tokens/s 894 ; Perplexity 10.93	
Epoch 13 ; Iteration 1300/1587 ; Learning rate 0.0625 ; Source tokens/s 891 ; Perplexity 10.91	
Epoch 13 ; Iteration 1350/1587 ; Learning rate 0.0625 ; Source tokens/s 890 ; Perplexity 10.95	
Epoch 13 ; Iteration 1400/1587 ; Learning rate 0.0625 ; Source tokens/s 888 ; Perplexity 10.96	
Epoch 13 ; Iteration 1450/1587 ; Learning rate 0.0625 ; Source tokens/s 886 ; Perplexity 10.95	
Epoch 13 ; Iteration 1500/1587 ; Learning rate 0.0625 ; Source tokens/s 884 ; Perplexity 10.95	
Epoch 13 ; Iteration 1550/1587 ; Learning rate 0.0625 ; Source tokens/s 883 ; Perplexity 10.94	
Validation perplexity: 22.913811798902	
Saving checkpoint to 'models/baseline_epoch13_22.91.t7'...