pytorch 加载 本地 roberta 模型

在利用Torch模块加载本地roberta模型时总是报OSERROR，如下：

OSError: Model name './chinese_roberta_wwm_ext_pytorch' was not found in tokenizers model name list (roberta-base, roberta-large, roberta-large-mnli, distilroberta-base, roberta-base-openai-detector, roberta-large-openai-detector). We assumed './chinese_roberta_wwm_ext_pytorch' was a path, a model identifier, or url to a directory containing vocabulary files named ['vocab.json', 'merges.txt'] but couldn't find such vocabulary files at this path or url.

解决方式：使用BertTokenizer以及BertModel加载，请勿使用RobertaTokenizer/RobertaModel，

如用RobertaForQuestionAnswering，如下的方式也可以，但是会有warning。

import torch
from transformers import BertTokenizer, BertModel,RobertaTokenizer,RobertaForQuestionAnswering
from transformers import logging

# logging.set_verbosity_warning()
# logging.set_verbosity_error()

tokenizer = BertTokenizer.from_pretrained("./chinese_roberta_wwm_ext_pytorch")
roberta = RobertaForQuestionAnswering.from_pretrained("./chinese_roberta_wwm_ext_pytorch",return_dict=True)

本文地址：https://blog.csdn.net/guyu1003/article/details/109024791