wwjCMP

> Hi， > > 请修改 `CUSTOM_API_KEY_PATTERN = r"^[A-Za-z0-9-_=]+\.[A-Za-z0-9-_=]+\.[A-Za-z0-9-_.+/=]*$"` > > 以适配 JWT 类型的 API_KEY API key现在可以了。我想问一下，怎么将openai的链接重定向到本地的服务。我在本地运行了下面链接的代理，可正常运行。但这里重定向失败。是设置的问题吗？ API_URL_REDIRECT: ' "https://api.openai.com/v1/chat/completions": "http://localhost:8006/v1/chat/completions" ' 

当然还可以增加一些常用的预处理选项，也可以提供正则表达式实现更加个性化的清洗

you had better put the model name into the title.

LANGFLOW_DATABASE_URL=sqlite:///./langflow.db LANGFLOW_LANGCHAIN_CACHE=SQLiteCache LANGFLOW_HOST=127.0.0.1 LANGFLOW_WORKERS=1 LANGFLOW_PORT=7866 LANGFLOW_LOG_LEVEL=critical LANGFLOW_LOG_FILE=logs/langflow.log LANGFLOW_FRONTEND_PATH=/path/to/frontend/build/files LANGFLOW_OPEN_BROWSER=true LANGFLOW_REMOVE_API_KEYS=false LANGFLOW_CACHE_TYPE=memory LANGFLOW_SUPERUSER=QAQ LANGFLOW_SUPERUSER_PASSWORD=123456 The above is my environment setup.

> > 回答结束，调用系统通知。在别的chat应用中用过，效果还不错。 > > 是回答结束后，让浏览器弹出提示吗？这好像有点点不必要，毕竟会话的等待时间不长。 不过KB的后台embedding倒是可以考虑可选的通知。 > > [useWebNotification](https://vueuse.org/core/useWebNotification/) 本地模型速度不快，需要等待一段时间

I want to know how to use slurm to run the Ollama service.

In fact, is supporting Gemini pro a better choice? Because it provides free api

Regarding your last question, I think the simplest way is that we don't need to completely hide the group; it can be folded into a column that occupies a small...

https://github.com/ollama/ollama/issues/4205

https://github.com/ggerganov/llama.cpp/issues/6803