batchtokens-8192_gpus-8.log

[valid] neckloss=0.10 codepp=9.34 lenloss=4.61 lenacc=0.01 loss=12.34 word_acc=0.00 kl=2.53 tokloss=7.40 tok+kl=9.93 * (epoch 1, step 1)
[TransformerScheduler] lr=0.0001
[valid] neckloss=0.02 codepp=6.56 lenloss=4.83 lenacc=0.00 loss=13.42 word_acc=0.12 kl=6.24 tokloss=4.74 tok+kl=10.98 (epoch 1, step 328)
[valid] neckloss=0.03 codepp=22.59 lenloss=4.65 lenacc=0.01 loss=11.20 word_acc=0.32 kl=6.10 tokloss=2.74 tok+kl=8.84 * (epoch 1, step 655)
[valid] neckloss=0.05 codepp=52.40 lenloss=4.63 lenacc=0.02 loss=10.68 word_acc=0.40 kl=6.19 tokloss=2.12 tok+kl=8.31 * (epoch 1, step 982)
[valid] neckloss=0.10 codepp=120.26 lenloss=4.67 lenacc=0.00 loss=10.07 word_acc=0.51 kl=6.15 tokloss=1.49 tok+kl=7.64 * (epoch 1, step 1309)
[train] neckloss=0.05 codepp=38.60 lenloss=4.77 lenacc=0.01 loss=12.16 word_acc=0.27 kl=6.27 tokloss=3.46 tok+kl=9.73
[nmtlab] Ending epoch 1, spent 15 minutes
[TransformerScheduler] lr=0.0005.9 token/s
[valid] neckloss=0.14 codepp=158.60 lenloss=4.62 lenacc=0.00 loss=9.73 word_acc=0.55 kl=6.11 tokloss=1.16 tok+kl=7.28 * (epoch 2, step 1638)
[valid] neckloss=0.19 codepp=181.43 lenloss=4.63 lenacc=0.01 loss=9.54 word_acc=0.60 kl=6.09 tokloss=0.93 tok+kl=7.03 * (epoch 2, step 1965)
[valid] neckloss=0.25 codepp=189.03 lenloss=4.63 lenacc=0.01 loss=9.45 word_acc=0.63 kl=6.06 tokloss=0.82 tok+kl=6.88 * (epoch 2, step 2292)
[valid] neckloss=0.32 codepp=196.67 lenloss=4.64 lenacc=0.01 loss=9.41 word_acc=0.65 kl=6.03 tokloss=0.73 tok+kl=6.76 * (epoch 2, step 2619)
[train] neckloss=0.21 codepp=178.17 lenloss=4.63 lenacc=0.01 loss=9.80 word_acc=0.56 kl=6.09 tokloss=1.18 tok+kl=7.27
[nmtlab] Ending epoch 2, spent 15 minutes
[TransformerScheduler] lr=0.00097 token/s
[valid] neckloss=0.42 codepp=194.05 lenloss=4.68 lenacc=0.01 loss=9.40 word_acc=0.66 kl=5.96 tokloss=0.68 tok+kl=6.64 * (epoch 3, step 2948)
[valid] neckloss=0.55 codepp=192.78 lenloss=4.72 lenacc=0.01 loss=9.45 word_acc=0.66 kl=5.88 tokloss=0.66 tok+kl=6.54 * (epoch 3, step 3275)
[valid] neckloss=0.70 codepp=194.47 lenloss=4.80 lenacc=0.01 loss=9.59 word_acc=0.66 kl=5.83 tokloss=0.66 tok+kl=6.49 * (epoch 3, step 3602)
[valid] neckloss=0.89 codepp=195.19 lenloss=4.91 lenacc=0.00 loss=9.75 word_acc=0.66 kl=5.76 tokloss=0.65 tok+kl=6.40 * (epoch 3, step 3929)
[train] neckloss=0.60 codepp=202.78 lenloss=4.73 lenacc=0.01 loss=9.74 word_acc=0.62 kl=5.90 tokloss=0.87 tok+kl=6.77
[nmtlab] Ending epoch 3, spent 15 minutes
[TransformerScheduler] lr=0.0014.0 token/s
[valid] neckloss=1.13 codepp=194.23 lenloss=4.91 lenacc=0.01 loss=9.97 word_acc=0.65 kl=5.70 tokloss=0.69 tok+kl=6.39 * (epoch 4, step 4258)
[valid] neckloss=1.38 codepp=196.70 lenloss=4.96 lenacc=0.00 loss=10.20 word_acc=0.65 kl=5.66 tokloss=0.68 tok+kl=6.34 * (epoch 4, step 4585)
[valid] neckloss=1.64 codepp=193.10 lenloss=5.10 lenacc=0.01 loss=10.54 word_acc=0.65 kl=5.65 tokloss=0.70 tok+kl=6.35 (epoch 4, step 4912)
[valid] neckloss=1.91 codepp=195.90 lenloss=5.10 lenacc=0.00 loss=10.80 word_acc=0.65 kl=5.63 tokloss=0.71 tok+kl=6.34 (epoch 4, step 5239)
[train] neckloss=1.48 codepp=202.64 lenloss=4.98 lenacc=0.00 loss=10.59 word_acc=0.61 kl=5.69 tokloss=0.93 tok+kl=6.61
[nmtlab] Ending epoch 4, spent 14 minutes