Writing logs to /p/qdata/jm8wx/research/text_attacks/textattack/outputs/training/distilbert-base-uncased-ag_news-2020-07-03-06:26/log.txt.
Loading nlp dataset ag_news, split train.
Loading nlp dataset ag_news, split test.
Loaded dataset. Found: 4 labels: ([0, 1, 2, 3])
Loading transformers AutoModelForSequenceClassification: distilbert-base-uncased
Tokenizing training data. (len: 120000)
Tokenizing eval data (len: 7600)
Loaded data and tokenized in 145.95597338676453s
Training model across 4 GPUs
***** Running training *****
	Num examples = 120000
	Batch size = 32
	Max sequence length = 128
	Num steps = 18750
	Num epochs = 5
	Learning rate = 2e-05
Eval accuracy: 93.94736842105263%
Best acc found. Saved model to /p/qdata/jm8wx/research/text_attacks/textattack/outputs/training/distilbert-base-uncased-ag_news-2020-07-03-06:26/.
Eval accuracy: 94.78947368421052%
Best acc found. Saved model to /p/qdata/jm8wx/research/text_attacks/textattack/outputs/training/distilbert-base-uncased-ag_news-2020-07-03-06:26/.
Eval accuracy: 94.67105263157895%
Eval accuracy: 94.72368421052632%
Eval accuracy: 94.56578947368422%
Saved tokenizer <textattack.models.tokenizers.auto_tokenizer.AutoTokenizer object at 0x7f914302fc10> to /p/qdata/jm8wx/research/text_attacks/textattack/outputs/training/distilbert-base-uncased-ag_news-2020-07-03-06:26/.
Wrote README to /p/qdata/jm8wx/research/text_attacks/textattack/outputs/training/distilbert-base-uncased-ag_news-2020-07-03-06:26/README.md.
Wrote training args to /p/qdata/jm8wx/research/text_attacks/textattack/outputs/training/distilbert-base-uncased-ag_news-2020-07-03-06:26/train_args.json.