运行generate.py报错

[louis-gg]

下载散文模板，在工程目录下建文件夹model/sanwen/将文件config.json、pytorch_model.bin、vocab.txt放入，修改源码： parser.add_argument( "--model_path", default="model/sanwen/", type=str, required=False, help="模型路径", ) 执行后报错：（已经确定文件目录存在） Traceback (most recent call last): File "./generate.py", line 233, in main() File "./generate.py", line 206, in main for key, value in torch.load(args.model_path, map_location="cpu")[ File "C:\ProgramData\Anaconda3\lib\site-packages\torch\serialization.py", line 579, in load with _open_file_like(f, 'rb') as opened_file: File "C:\ProgramData\Anaconda3\lib\site-packages\torch\serialization.py", line 230, in _open_file_like return _open_file(name_or_buffer, mode) File "C:\ProgramData\Anaconda3\lib\site-packages\torch\serialization.py", line 211, in init super(_open_file, self).init(open(name, mode)) PermissionError: [Errno 13] Permission denied: 'model/sanwen/'

[Lirsakura]

请问一下你的训练代码是最新的吗，出现过attention_mask报错吗

[Creative-Ataraxia]

会不会是model/sanwen/没有read access