feat: set default model for image description #4687

Closed
opened 2025-11-11 16:00:40 -06:00 by GiteaMirror · 2 comments
Owner

Originally created by @space192 on GitHub (Apr 3, 2025).

Check Existing Issues

  • I have searched the existing issues and discussions.

Problem Description

when chatting with a LLM with no vision capabilities, switching to a vision LLM to describe the image can create issue due to the previous context, could be nice to auto switch to just describe the image

Desired Solution you'd like

The idea would be that the user / admin can set a default model for image understanding such that when chatting with a LLM with no vision capabilities, if an image is submitted by the user it automatically described by the choosed vision LLM and then the original LLM continues with it's work

Alternatives Considered

create a tool where the image is sent to the other LLM

Additional Context

No response

Originally created by @space192 on GitHub (Apr 3, 2025). ### Check Existing Issues - [x] I have searched the existing issues and discussions. ### Problem Description when chatting with a LLM with no vision capabilities, switching to a vision LLM to describe the image can create issue due to the previous context, could be nice to auto switch to just describe the image ### Desired Solution you'd like The idea would be that the user / admin can set a default model for image understanding such that when chatting with a LLM with no vision capabilities, if an image is submitted by the user it automatically described by the choosed vision LLM and then the original LLM continues with it's work ### Alternatives Considered create a tool where the image is sent to the other LLM ### Additional Context _No response_
Author
Owner

@spammenotinoz commented on GitHub (Apr 3, 2025):

Another option is not to send to the LLM at all.
This might be a feature to hold-off and see where she lands. LLM models now are starting to generate and modify their own images. So perhaps a separate image model \ function no longer makes as much sense as it did previously.

@spammenotinoz commented on GitHub (Apr 3, 2025): Another option is not to send to the LLM at all. This might be a feature to hold-off and see where she lands. LLM models now are starting to generate and modify their own images. So perhaps a separate image model \ function no longer makes as much sense as it did previously.
Author
Owner

@tjbck commented on GitHub (Apr 3, 2025):

PLEASE check for existing issues and discussions first, this can be implemented via Functions.

@tjbck commented on GitHub (Apr 3, 2025): PLEASE check for existing issues and discussions first, this can be implemented via Functions.
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/open-webui#4687