Avoid contention when loading models

[Btlmd]

Model loading is very likely to trigger this error

Traceback (most recent call last):    
  File "main.py", line 434, in <module>                                                                                                                                       
    main()                                                                                                                                                                    
  File "main.py", line 114, in main    
    tokenizer = AutoTokenizer.from_pretrained(model_args.model_name_or_path, trust_remote_code=True)                                                                          
  File "/opt/conda/lib/python3.8/site-packages/transformers/models/auto/tokenization_auto.py", line 663, in from_pretrained                                                   
    tokenizer_class = get_class_from_dynamic_module(                                   
  File "/opt/conda/lib/python3.8/site-packages/transformers/dynamic_module_utils.py", line 388, in get_class_from_dynamic_module                                              
    final_module = get_cached_module_file(                                                                                                                                    
  File "/opt/conda/lib/python3.8/site-packages/transformers/dynamic_module_utils.py", line 279, in get_cached_module_file                                                     
    shutil.copy(resolved_module_file, submodule_path / module_file)                                                                                                           
  File "/opt/conda/lib/python3.8/shutil.py", line 419, in copy                                                                                                                
    copymode(src, dst, follow_symlinks=follow_symlinks)                                
  File "/opt/conda/lib/python3.8/shutil.py", line 308, in copymode                                                                                                            
    chmod_func(dst, stat.S_IMODE(st.st_mode))                                          
FileNotFoundError: [Errno 2] No such file or directory: '/root/.cache/huggingface/modules/transformers_modules/chatglm-6b/tokenization_chatglm.py' 

because the model are loaded simultaneoutly in each process.

Add a lock to prevent this.

[roki1031]

我按照您的代码在本地做了修改，还是会报一样的错误。