error-handling - 如何用pytorch实现SciBERT；加载时出错

Question

我正在尝试使用SciBERT预先训练的模型，即:scibert-scivocab-uncased以下方式:

    !pip install pytorch-pretrained-bert 
    import torch
    from pytorch_pretrained_bert import BertTokenizer, BertModel,      BertForMaskedLM 
    import logging
    import matplotlib.pyplot as plt
    tokenizer = BertTokenizer.from_pretrained('bert-base-uncased')
    indexed_tokens = tokenizer.convert_tokens_to_ids(tokenized_text) 
    segments_ids = [1] * len(tokenized_text)
    tokens_tensor = torch.tensor([indexed_tokens])
    segments_tensors = torch.tensor([segments_ids])
    model =      BertModel.from_pretrained('/Users/.../Downloads/scibert_scivocab_uncased-3.tar.gz') 

我收到以下错误:

EOFError: Compressed file ended before the end-of-stream marker was reached

我从网站(https://github.com/allenai/scibert)上下载了文件

我将其从“tar”转换为gzip

没事。

关于如何处理此问题的任何提示？

谢谢!

Answer 1

在新版本的pytorch-pretrained-BERT中，即在变压器中，您可以执行以下操作以在解压缩后加载预训练的模型:

导入AutoModelForTokenClassification，AutoTokenizer

模型= AutoModelForTokenClassification.from_pretrained(“/您的/本地/路径/到/scibert_scivocab_uncased”)