Writing logs to ./outputs/2024-03-01-08-56-39-463700/train_log.txt.
Wrote original training args to ./outputs/2024-03-01-08-56-39-463700/training_args.json.
***** Running training *****
  Num examples = 50000
  Num epochs = 3
  Num clean epochs = 3
  Instantaneous batch size per device = 8
  Total train batch size (w. parallel, distributed & accumulation) = 8
  Gradient accumulation steps = 1
  Total optimization steps = 18750
==========================================================
Epoch 1
Running clean epoch 1/3
Train accuracy: 83.47%
Eval accuracy: 86.99%
Best score found. Saved model to ./outputs/2024-03-01-08-56-39-463700/best_model/
==========================================================
Epoch 2
Running clean epoch 2/3
Train accuracy: 90.54%
Eval accuracy: 89.44%
Best score found. Saved model to ./outputs/2024-03-01-08-56-39-463700/best_model/
==========================================================
Epoch 3
Running clean epoch 3/3
Train accuracy: 95.00%
Eval accuracy: 90.22%
Best score found. Saved model to ./outputs/2024-03-01-08-56-39-463700/best_model/
Wrote README to ./outputs/2024-03-01-08-56-39-463700/README.md.