riverbutte
riverbutte
@chaideblog i met with the same problem. Try to copy all the codes in 'find_mxnet.py' to 'visualize_net.py', and delete the first two lines.
@Mrxiexianzhao 自己写得了，反正有 dmPython 可以连接达梦
Solve it ! https://github.com/eosphoros-ai/DB-GPT/issues/1604#issuecomment-2151499049
Looking forward to great solution. This feature is indeed very necessary because according to the provided QA file, users can directly match Q-Q when asking questions, which is very beneficial...
LLM_MODEL=proxyllm PROXY_SERVER_URL=http://172.16.15.242:8000/v1/chat/completions PROXYLLM_BACKEND=qwen2_5-14b-instruct
Second that! Has the community considered separating the API and task_executor? The API supports load balancing, while task_executor supports horizontal scaling?
@KevinHuSh > Yes, it can. Where can i see it? Even if the latest version is deployed locally, I haven't seen the effect of mixing images and texts to answer.