Kevin Hu comments

Results 1512 comments of


                                            Kevin Hu

[Bug]: Exception: Can't connect to ES cluster

**`docker-compose-gpu-CN-oc9`** is not maintained any more.

[Bug]: Exception: Can't connect to ES cluster

[FYI](https://docs.docker.com/compose/how-tos/gpu-support/)

[Question]: Multi-turn optimization open，After multiple rounds of dialogue, regardless of what questions are asked subsequently, an empty response message will be returned directly

You could click the little lamp to have a look what exactly question changed to. ![Image](https://github.com/user-attachments/assets/3ad5ffb8-616a-40d8-b23a-eeb84f869688)

[Question]: Multi-turn optimization open，After multiple rounds of dialogue, regardless of what questions are asked subsequently, an empty response message will be returned directly

Clean this field. ![Image](https://github.com/user-attachments/assets/99969741-885c-46e6-a01f-13d69ef2d7b6)

[Question]: RAGFlow 16 version Agent share link chat profile picture customization

Not supported yet.

[Question]: RAGFlow 16 version Agent share link chat profile picture customization

Ollama response error. Could paste the ollama error log here?

[Question]: RAGFlow 16 version Agent share link chat profile picture customization

I did not see any errors from it. Weird!...

[Bug]: Retrieval module truncates input prompts to 2048 tokens when using Ollama as Inference Provider

Weired! Did you set 16384 here? ![Image](https://github.com/user-attachments/assets/9c244095-0bf8-4583-9581-b6b797ab221b) And how do you know the prompt had been truncated?

[Question]: How do I use the built-in models of Ragflow?

You could search `BAAI/bge-large-zh-v1.5` in the codebase, and change it to the embedding model you want. But I suggest to serve embedding model via vLLM/Ollama/XInference independently.

[Question]: How do I use the built-in models of Ragflow?

What about signing up a new user?