hiyouga/LLaMA-Factory

api部署,method not allowed问题

jy-101361-1810897 opened this issue · 1 comments

Reminder

  • I have read the README and searched the existing issues.

Reproduction

CUDA_VISIBLE_DEVICES=1 API_PORT=9018 python src/api_demo.py
--model_name_or_path $base_model_dir
--adapter_name_or_path $lora_dir
--template chatglm3
--finetuning_type lora

Expected behavior

运行sh文件,在服务器上成功加载模型,vscode转发端口,在服务器/本地都无法成功访问地址,curl http://0.0.0.0:9018/v1/chat/completions出现method not allowed

System Info

No response

Others

No response

应该使用 post,而不是 get

https://zhuanlan.zhihu.com/p/695287607 的最后有调用例子