File size: 1,743 Bytes
6698e32 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 |
Writing logs to /p/qdata/jm8wx/research/text_attacks/textattack/outputs/training/distilbert-base-uncased-imdb-2020-06-30-09:57/log.txt. Loading [94mnlp[0m dataset [94mimdb[0m, split [94mtrain[0m. Loading [94mnlp[0m dataset [94mimdb[0m, split [94mtest[0m. Loaded dataset. Found: 2 labels: ([0, 1]) Loading transformers AutoModelForSequenceClassification: distilbert-base-uncased Tokenizing training data. (len: 25000) Tokenizing eval data (len: 25000) Loaded data and tokenized in 76.20483636856079s Training model across 4 GPUs ***** Running training ***** Num examples = 25000 Batch size = 16 Max sequence length = 128 Num steps = 7810 Num epochs = 5 Learning rate = 2e-05 Eval accuracy: 87.312% Best acc found. Saved model to /p/qdata/jm8wx/research/text_attacks/textattack/outputs/training/distilbert-base-uncased-imdb-2020-06-30-09:57/. Eval accuracy: 87.336% Best acc found. Saved model to /p/qdata/jm8wx/research/text_attacks/textattack/outputs/training/distilbert-base-uncased-imdb-2020-06-30-09:57/. Eval accuracy: 88.0% Best acc found. Saved model to /p/qdata/jm8wx/research/text_attacks/textattack/outputs/training/distilbert-base-uncased-imdb-2020-06-30-09:57/. Eval accuracy: 87.832% Eval accuracy: 87.736% Saved tokenizer <textattack.models.tokenizers.auto_tokenizer.AutoTokenizer object at 0x7f2581bcb400> to /p/qdata/jm8wx/research/text_attacks/textattack/outputs/training/distilbert-base-uncased-imdb-2020-06-30-09:57/. Wrote README to /p/qdata/jm8wx/research/text_attacks/textattack/outputs/training/distilbert-base-uncased-imdb-2020-06-30-09:57/README.md. Wrote training args to /p/qdata/jm8wx/research/text_attacks/textattack/outputs/training/distilbert-base-uncased-imdb-2020-06-30-09:57/train_args.json. |