| This model is developed with transformers v4.10.3. | |
| # Train | |
| ```bash | |
| #!/usr/bin/env bash | |
| export CUDA_VISIBLE_DEVICES=0 | |
| OUTDIR=bert-based-uncased-mnli | |
| WORKDIR=transformers/examples/pytorch/text-classification | |
| cd $WORKDIR | |
| nohup python run_glue.py \ | |
| --model_name_or_path bert-base-uncased \ | |
| --task_name mnli \ | |
| --do_eval \ | |
| --do_train \ | |
| --per_device_train_batch_size 16 \ | |
| --per_device_eval_batch_size 16 \ | |
| --max_seq_length 128 \ | |
| --num_train_epochs 3 \ | |
| --overwrite_output_dir \ | |
| --output_dir $OUTDIR 2>&1 | tee $OUTDIR/run.log & | |
| ``` | |
| # Eval | |
| ```bash | |
| export CUDA_VISIBLE_DEVICES=0 | |
| OUTDIR=eval-bert-based-uncased-mnli | |
| WORKDIR=transformers/examples/pytorch/text-classification | |
| cd $WORKDIR | |
| nohup python run_glue.py \ | |
| --model_name_or_path vuiseng9/bert-base-uncased-mnli \ | |
| --task_name mnli \ | |
| --do_eval \ | |
| --per_device_eval_batch_size 16 \ | |
| --max_seq_length 128 \ | |
| --overwrite_output_dir \ | |
| --output_dir $OUTDIR 2>&1 | tee $OUTDIR/run.log & | |
| ``` | |