trace 'THUDM/chatglm-6b-int8' 的时候报错：Could not export Python function call 'W8A16Linear'. Remove calls to Python functions before export.

THUDM / ChatGLM-6B

ChatGLM-6B: An Open Bilingual Dialogue Language Model | 开源双语对话语言模型

Apache License 2.0

40.03k stars 5.16k forks source link

trace 'THUDM/chatglm-6b-int8' 的时候报错：Could not export Python function call 'W8A16Linear'. Remove calls to Python functions before export. #1374

Open mymagicpower opened 11 months ago

mymagicpower commented 11 months ago

Is there an existing issue for this?

[X] I have searched the existing issues

Current Behavior

对于量化的模型 'THUDM/chatglm-6b-int8'，torch.jit.trace 的时候会报错。 Could not export Python function call 'W8A16Linear'. Remove calls to Python functions before export.

Expected Behavior

能正常，trace 出 'THUDM/chatglm-6b-int8' 模型。

Steps To Reproduce

torch.jit.trace(xxx,xxx)

Environment

这个问题与环境无关。

Anything else?

No response