openai server非流式请求报错 about chatglm.cpp HOT 6 CLOSED

启动命令：

(venv) C:\Windows\System32>uvicorn chatglm_cpp.openai_api:app --host 127.0.0.1 --port 4523

from chatglm.cpp.

I ran it on m2 ultra and encountered the same error.
The running command was "MODEL=./chatglm.cpp/models/chatglm4-ggml.bin uvicorn chatglm_cpp.openai_api:app --host 0.0.0.0 --port 8848"

from chatglm.cpp.

Fixed by #309. Will release a new version soon.

from chatglm.cpp.

I ran it on m2 ultra and encountered the same error. The running command was "MODEL=./chatglm.cpp/models/chatglm4-ggml.bin uvicorn chatglm_cpp.openai_api:app --host 0.0.0.0 --port 8848"

我查看了#309的修改方式，直接修改了site-packages中的代码，把encode_messages修改为apply_chat_template，不再报错了

from chatglm.cpp.

I ran it on m2 ultra and encountered the same error. The running command was "MODEL=./chatglm.cpp/models/chatglm4-ggml.bin uvicorn chatglm_cpp.openai_api:app --host 0.0.0.0 --port 8848"

我查看了#309的修改方式，直接修改了site-packages中的代码，把encode_messages修改为apply_chat_template，不再报错了

我暂时也用相同的方式解决了，多谢！

from chatglm.cpp.

已经发布新版本了，可以更新下 python 包，先关了

from chatglm.cpp.