do_predict问题

[Yesgo1220]

对torch不太熟悉，请问在只do_predict的时候，没有找到load_state_dict这样的操作，模型是怎么预测的？麻烦指点一二，感谢。作者是否应该在do_predict中加上model.load_state_dict(torch.load(model_name))

[liuwei1206]

Hi,

Thanks for your comment. When do evaluation or prediction, I will change the model_name_or_path with the trained checkpoint. In this way, you can load it with from_pretrain function.

Best