您好，非常感谢您的工作。我进行全参数微调后，调用infer_finetuning.py的时候，出现Missing key(s) in state_dict: "_TransformerLightningModule__backbone.transformer.lm_head.weight".请问您有遇到过这个问题吗

[ssbuild]

https://github.com/ssbuild/chatglm_finetuning/blob/089404c0ffd59ea5c9e2d889ccb4c48f866c867c/infer_finetuning.py#L40 lm_head 是共享权重 ，修改 strict = False 即可

[MrTFK]

我这边修改strict=False后输出的结果是很差的，基本等同乱码，是不是不太能这么搞，感觉就像参数错乱了

比如： input 影响黄金价格的因素有哪些？ output 等等。等等，等等，等等，等等，等等，等等，等等，等等，等等，等等，等等，等等，等等，等等，等等，等等，等等，等等，等等，等等，等等，等等，等等，等等，等等，等等，等等，等等，等等，等等，474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747474747