Can not train with a single GPU

[CHAOS-Yang]

I run the code with

CUDA_VISIBLE_DEVICES=0 python -m classification.run_wrapper --output_dir ./classification/checkpoints --task_name sst-2 --target_epsilon 1 --model_name_or_path bert-base-uncased --data_dir ./classification/data

and get the result:

src/trainer.py", line 254, in train raise ValueError("Multi-gpu and distributed training is currently not supported.") ValueError: Multi-gpu and distributed training is currently not supported.

I try all the args and can not disable the Mutil-GPU training?

[CHAOS-Yang]

I find the problem that should keep transformers==4.20.1