/
rerunW2V.sh
19 lines (19 loc) · 2.5 KB
/
rerunW2V.sh
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
jmoreno@moriche:~/Documents/Projects/MultiEmbeddings/embedding$ python3 word2vec.py --train_data=wiki_all.clean2 --eval_data=questions-words.txt --save_path=./ --epochs_to_train=5 --checkpoint_interval=10800
2017-11-23 13:25:45.754755: I tensorflow/core/platform/cpu_feature_guard.cc:137] Your CPU supports instructions that this TensorFlow binary was not compiled to use: SSE4.1 SSE4.2 AVX AVX2 FMA
2017-11-23 13:31:45.489638: I word2vec_kernels.cc:211] Data file: wiki_all.clean2 contains 14967299136 bytes, 2150648153 words, 30813400 unique words, 5245842 unique frequent words.
Data file: wiki_all.clean2
Vocab size: 5245842 + UNK
Words per epoch: 2150648153
Eval analogy file: questions-words.txt
Questions: 19544
Skipped: 0
Epoch 1 Step 555374495: lr = 0.172 loss = 0.17 words/sec = 22076
Eval 7929/19544 accuracy = 40.6%
Epoch 2 Step 1110702795: lr = 0.145 loss = 0.77 words/sec = 16845
Eval 9166/19544 accuracy = 46.9%
Epoch 2 Step 1363801946: lr = 0.132 loss = 5.63 words/sec = 2165Epoch 2 Step 1363842814: lr = 0.132 loss = 3.34 words/sec = 2179Epoch 2 Step 1363883724: lr = 0.132 loss = 4.93 words/sec = 2183Epoch 2 Step 1363924217: lr = 0.132 loss = 4.44 words/sec = 2163Epoch 2 Step 1363965019: lr = 0.132 loss = 4.37 words/sec = 2179Epoch 2 Step 1364005952: lr = 0.132 loss = 4.41 words/sec = 2180Epoch 2 Step 1364046779: lr = 0.132 loss = 4.86 words/sec = 2182Epoch 2 Step 1364087790: lr = 0.132 loss = 4.14 words/sec = 2194Epoch 2 Step 1364128484: lr = 0.132 loss = 5.00 words/sec = 2168Epoch 2 Step 1364169123: lr = 0.132 loss = 3.67 words/sec = 2165Epoch 2 Step 1364209893: lr = 0.132 loss = 4.81 words/sec = 2176Epoch 2 Step 1364250982: lr = 0.132 loss = 4.67 words/sec = 2190Epoch 2 Step 1364291908: lr = 0.132 loss = 4.36 words/sec = 2174Epoch 2 Step 1364332787: lr = 0.132 loss = 5.41 words/sec = 2196Epoch 2 Step 1364373749: lr = 0.132 loss = 3.93 words/sec = 2177Epoch 2 Step 1364414314: lr = 0.132 loss = 2.86 words/sec = 2172Epoch 2 Step 1364455213: lr = 0.132 loss = 4.22 words/sec = 2178Epoch 2 Step 1364495783: lr = 0.132 losEpoch 3 Step 1666035101: lr = 0.117 loss = 0.32 words/sec = 13890sec = 21813
Eval 10366/19544 accuracy = 53.0%
Epoch 4 Step -2073594872: lr = 0.089 loss = 0.26 words/sec = 1698
Eval 11282/19544 accuracy = 57.7%
Epoch 5 Step -1518246379: lr = 0.062 loss = 0.21 words/sec = 17443
Eval 12116/19544 accuracy = 62.0%