从BERT换成此模型后跑不通，应该怎么解决，显示缺少参数

[ottt08]

_from_pretrained() missing 1 required positional argument: 'init_configuration'
File "C:\Users\10553\Desktop\chinese_ner-main\torch_ner\source\ner_main.py", line 63, in train
tokenizer = BertTokenizer.from_pretrained("chinese-roberta-wwm-ext-large", do_lower_case=self.config.do_lower_case)
File "C:\Users\10553\Desktop\chinese_ner-main\torch_ner\source\ner_main.py", line 258, in
NerMain().train()

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.