shunjiu

使用Wenzhong2.0-GPT2-3.5B微调后生成乱码

9

你好，我使用Wenzhong2.0-GPT2-3.5B在下游任务微调后，预测结果是这种乱码，请问有解决办法吗？谢谢🙏 预测生成代码： tokenizer = GPT2Tokenizer.from_pretrained(model_path) model = GPT2LMHeadModel.from_pretrained(model_path) generator = pipeline('text-generation', model=model, tokenizer=tokenizer) print(generator(context, max_length=100, num_return_sequences=1)) 根据context生成结果： grinning pres ausp grinning grinning pres spectator Romo Wad Wad grinning 283 intellectual...