JSONDecodeError: Unexpected UTF-8 BOM (decode using utf-8-sig): line 1 column 1 (char 0) ---While Tuning gpt2.finetune

问题描述

希望你们一切都好， 我正在微调 GPT 2 模型以根据内容生成标题，在处理它时，我创建了一个仅包含标题的简单 CSV 文件来训练模型，但是在将此模型输入到 GPT 2 以进行微调时，我收到以下错误， JSONDecodeError 追溯（最近一次调用最后一次） 在 （） 10步=1000， 11 save_every=200,---> 12 sample_every=25) # steps 是最大训练步骤数 13 14 # gpt2.generate(sess)

    3 frames
    /usr/lib/python3.7/json/__init__.py in loads(s,encoding,cls,object_hook,parse_float,parse_int,parse_constant,object_pairs_hook,**kw)
        336         if s.startswith('\ufeff'):
        337           s = s.encode('utf8')[3:].decode('utf8')
    --> 338             # raise JSONDecodeError("Unexpected UTF-8 BOM (decode using utf-8-sig)",339             #                       s,0)
        340     else:
    
    JSONDecodeError: Unexpected UTF-8 BOM (decode using utf-8-sig): line 1 column 1 (char 0)
    
    Below is my code for the above :
    
    import gpt_2_simple as gpt2
    
    model_name = "120M" # "355M" for larger model (it's 1.4 GB)
    gpt2.download_gpt2(model_name=model_name)   # model is saved into current directory under /models/117M/
    sess = gpt2.start_tf_sess()
    
    gpt2.finetune(sess,'titles.csv',model_name=model_name,steps=1000,save_every=200,sample_every=25)   # steps is max number of training steps
    
    I have tried all the basic mechanism of handing UTF -8 BOM but did not find any luck,Hence requesting your help .It would be a great help from you all .

解决方法

尝试更改型号名称，因为我看到您输入了 120M 而 gpt2 型号称为 124M