RuntimeError: device type error

[Chen9154]

I copy the training command from text match tutorial and modify a little.

python -m torch.distributed.launch $DISTRIBUTED_ARGS main.py \ --mode train \ --worker_gpu=1 \ --tables=train.csv,dev.csv \ --input_schema=example_id:str:1,sent1:str:1,sent2:str:1,label:str:1,cate:str:1,score:str:1 \ --first_sequence=sent1 \ --second_sequence=sent2 \ --label_name=label \ --label_enumerate_values=0,1 \ --checkpoint_dir=./text_match_two_tower_model_dir \ --learning_rate=3e-5 \ --epoch_num=1 \ --random_seed=42 \ --save_checkpoint_steps=100 \ --sequence_length=128 \ --train_batch_size=32 \ --app_name=text_match \ --user_defined_parameters=' pretrain_model_name_or_path=hfl/chinese-roberta-wwm-ext two_tower=True loss_type=hinge_loss margin=0.45 gamma=32 embedding_size=256 '

(the code in knowledge_language_understanding shares the same error.

And here's the error stack in text_match.

[Chen9154]

fixed by replacing "gpu:0" with "cuda:0" in EasyNLP/easynlp/appzoo/text_match/evaluator.py.