vllm load model error
#3
by
Dharma0818
- opened
When I use vllm to load the gemma:2-27b-chinese-chat model, I receive the following error message. What is causing this issue?
ERROR 07-12 10:40:51 serving_chat.py:238] Error in applying chat template from request: 'system_message' is undefined
INFO: 100.89.61.1:55874 - "POST /v1/chat/completions HTTP/1.1" 400 Bad Request
the same situation in xinference "'system_message' is undefined"
在 xinference 加速推理框架中，同样提示“'system_message' is undefined”