Writing logs to /p/qdata/jm8wx/research/text_attacks/textattack/outputs/training/albert-base-v2-glue:wnli-2020-06-29-11:21/log.txt. Loading nlp dataset glue, subset wnli, split train. Loading nlp dataset glue, subset wnli, split validation. Loaded dataset. Found: 2 labels: ([0, 1]) Loading transformers AutoModelForSequenceClassification: albert-base-v2 Tokenizing training data. (len: 635) Tokenizing eval data (len: 71) Loaded data and tokenized in 4.413618564605713s Training model across 4 GPUs ***** Running training ***** Num examples = 635 Batch size = 64 Max sequence length = 256 Num steps = 45 Num epochs = 5 Learning rate = 2e-05 Eval accuracy: 59.154929577464785% Best acc found. Saved model to /p/qdata/jm8wx/research/text_attacks/textattack/outputs/training/albert-base-v2-glue:wnli-2020-06-29-11:21/. Eval accuracy: 47.88732394366197% Eval accuracy: 45.07042253521127% Eval accuracy: 47.88732394366197% Eval accuracy: 50.70422535211267% Saved tokenizer to /p/qdata/jm8wx/research/text_attacks/textattack/outputs/training/albert-base-v2-glue:wnli-2020-06-29-11:21/. Wrote README to /p/qdata/jm8wx/research/text_attacks/textattack/outputs/training/albert-base-v2-glue:wnli-2020-06-29-11:21/README.md. Wrote training args to /p/qdata/jm8wx/research/text_attacks/textattack/outputs/training/albert-base-v2-glue:wnli-2020-06-29-11:21/train_args.json. Writing logs to /p/qdata/jm8wx/research/text_attacks/textattack/outputs/training/albert-base-v2-glue:wnli-2020-06-29-11:21/log.txt. Loading nlp dataset glue, subset wnli, split train. Loading nlp dataset glue, subset wnli, split validation. Loaded dataset. Found: 2 labels: ([0, 1]) Loading transformers AutoModelForSequenceClassification: albert-base-v2 Tokenizing training data. (len: 635) Tokenizing eval data (len: 71) Loaded data and tokenized in 4.476848840713501s Training model across 4 GPUs ***** Running training ***** Num examples = 635 Batch size = 128 Max sequence length = 256 Num steps = 20 Num epochs = 5 Learning rate = 2e-05