[Question]: 信息抽取用的finetune.py运行报错

### 请提出你的问题

在线下调用PaddleNLP\PaddleNLP\applications\information_extraction\document\finetune.py 的时候，paddlenlp版本2.5.1 paddle 报错
---------------------------------------------------------------------------
UnboundLocalError                         Traceback (most recent call last)
File E:\0926test\大模型训练\PaddleNLP\PaddleNLP\applications\information_extraction\document\finetune.py:177
    173         export_model(model=trainer.model, input_spec=input_spec, path=model_args.export_model_dir)
    176 if __name__ == "__main__":
--> 177     main()

File E:\0926test\大模型训练\PaddleNLP\PaddleNLP\applications\information_extraction\document\finetune.py:115, in main()
    112 tokenizer = AutoTokenizer.from_pretrained(model_args.model_name_or_path)
    114 # Load and preprocess dataset
--> 115 train_ds = load_dataset(reader, data_path=data_args.train_path, max_seq_len=data_args.max_seq_len, lazy=False)
    116 dev_ds = load_dataset(reader, data_path=data_args.dev_path, max_seq_len=data_args.max_seq_len, lazy=False)
    117 trans_fn = partial(
    118     convert_example,
    119     tokenizer=tokenizer,
    120     max_seq_len=data_args.max_seq_len,
    121     dynamic_max_length=data_args.dynamic_max_length,
    122 )

File E:\Anaconda3\envs\pytorch\lib\site-packages\paddlenlp\datasets\dataset.py:186, in load_dataset(path_or_read_func, name, data_files, splits, lazy, **kwargs)
    183             custom_kwargs[name] = kwargs[name]
    185     reader_instance = SimpleBuilder(lazy=lazy, read_func=path_or_read_func)
--> 186     return reader_instance.read(**custom_kwargs)
    187 else:
    188     try:

File E:\Anaconda3\envs\pytorch\lib\site-packages\paddlenlp\datasets\dataset.py:768, in SimpleBuilder.read(self, **kwargs)
    766     return MapDataset(examples)
    767 else:
--> 768     return MapDataset(list(examples))

File E:\0926test\大模型训练\PaddleNLP\PaddleNLP\applications\information_extraction\document\utils.py:110, in reader(data_path, max_seq_len)
    108     else:
    109         content = res_content
--> 110         boxes = res_boxes
    112 for json_line in json_lines:
    113     yield json_line

UnboundLocalError: local variable 'res_boxes' referenced before assignmen

PaddlePaddle/PaddleNLP

内容风险标识

评论 (0)

PaddlePaddle/PaddleNLP .gitee-modal { width: 500px !important; }

内容风险标识