报错Can't load the configuration of WisdomShell--CodeShell-7B-Chat\config.json
big-F opened this issue · comments
我是直接在huggingface上把文件下到本地了。下的是这里的文件
https://huggingface.co/WisdomShell/CodeShell-7B-Chat/tree/main
然后我在web_demo.py做了修改
def _load_model_tokenizer(args):
tokenizer = AutoTokenizer.from_pretrained(
# args.checkpoint_path, trust_remote_code=True, resume_download=True,
'D:\models\WisdomShell--CodeShell-7B-Chat',
)
model = AutoModelForCausalLM.from_pretrained(
# args.checkpoint_path,
'D:\\models\\WisdomShell--CodeShell-7B-Chat',
device_map=args.device,
trust_remote_code=True,
resume_download=True,
torch_dtype=torch.bfloat16,
).eval()
config = GenerationConfig.from_pretrained(
# args.checkpoint_path, trust_remote_code=True, resume_download=True,
'D:\\models\\WisdomShell--CodeShell-7B-Chat\\config.json',
)
运行报错
OSError: Can't load the configuration of 'D:\models\WisdomShell--CodeShell-7B-Chat\config.json'. If you were trying to load it from 'https://huggingface.co/models', make sure you don't have a local directory with the same name. Otherwise, make sure 'D:\models\WisdomShell--CodeShell-7B-Chat\config.json' is the correct path to a directory containing a generation_config.json file
我看config.json里的"_name_or_path": "/nvme/xr/checkpoints/codeshell/pt_codeshell/iter_0023208/hf",是这样的
请问我该如何解决,谢谢
config = GenerationConfig.from_pretrained(
# args.checkpoint_path, trust_remote_code=True, resume_download=True,
'D:\models\WisdomShell--CodeShell-7B-Chat',
这样就可以运行了。
https://huggingface.co/docs/transformers/installation误导了我