Unable to load weights from pytorch checkpoint file.

[bixingjian]

Describe the question

A clear and concise description of what the question is. 运行lm的时候出现的问题. 在git clone https://huggingface.co/hfl/chinese-bert-wwm的文件放到了pretrained文件夹下面.

# 当使用预训练语言模型时，该预训练的模型存放位置
# lm_name = 'bert-base-chinese'  # download usage
lm_file: '/data/my_user_name/deepke/pretrained/chinese-bert-wwm'
# lm_file: '/home/yhy/transformers/bert-base-chinese'

请问bert模型应该下载哪个文件, 并且config文件如何配置呢.

Environment (please complete the following information):

• OS: Ubuntu
• Python Version: py36
• torch: 1.4

Screenshots

If applicable, add screenshots to help explain your problem.

Additional context

Add any other context about the problem here.

[bixingjian]

Just something wrong with downloading the pretrained files, remember to use git-lfs.