Use Bert pre-trained model to fine-tune a binary text classifier with
run_classifier.py --task_name=my_processor --do_train=true --do_eval=true --data_dir=$BERT/DATA/DIR/data --vocab_file=$BERT_BASE_DIR/vocab.txt --bert_config_file=$BERT_BASE_DIR/bert_config.json --init_checkpoint=$BERT_BASE_DIR/bert_model.ckpt --max_seq_length=128 --train_batch_size=32 --learning_rate=2e-5 --num_train_epochs=3.0 --output_dir=./result/
*** Eval Results ***
| eval_accuracy | eval_loss |
|---|---|
| 0.9985759 | 0.008041907 |