[GH-ISSUE #16768] feat: issue: title / tag generation offload current model #33567

Closed
opened 2026-04-25 07:27:53 -05:00 by GiteaMirror · 1 comment
Owner

Originally created by @bchtrue on GitHub (Aug 21, 2025).
Original GitHub issue: https://github.com/open-webui/open-webui/issues/16768

Actual Behavior

Loading the model is time-consuming, and tasks such as title generation, tags, and follow-ups offload the "current model".

While I have adjusted the settings to use the "current" model for tags and title generation still offloads the "current" model due to differing context window size.

Suggestion:
I propose adding a setting that allows the title generation feature to utilize the same context window as previous request if the setting is set to "current model" or add settings of context window size near model choose for tags/titles/follow up(currently don't have) generation. This would enhance efficiency by reducing offloading of current model, loading new model, offloading it and loading first one for next request.

Thank you for considering this enhancement!

Originally created by @bchtrue on GitHub (Aug 21, 2025). Original GitHub issue: https://github.com/open-webui/open-webui/issues/16768 ### Actual Behavior Loading the model is time-consuming, and tasks such as title generation, tags, and follow-ups offload the "current model". While I have adjusted the settings to use the "current" model for tags and title generation still offloads the "current" model **due to differing context window size.** **Suggestion:** I propose adding a setting that allows the title generation feature to utilize **the same context window as previous request** if the setting is set to "current model" or add settings of context window size near model choose for tags/titles/follow up(currently don't have) generation. This would enhance efficiency by reducing offloading of current model, loading new model, offloading it and loading first one for next request. Thank you for considering this enhancement!
GiteaMirror added the bug label 2026-04-25 07:27:53 -05:00
Author
Owner

@tjbck commented on GitHub (Aug 21, 2025):

This should already be the case, could you confirm you've modified the context length directly from the model editor?

<!-- gh-comment-id:3209600222 --> @tjbck commented on GitHub (Aug 21, 2025): This should already be the case, could you confirm you've modified the context length directly from the model editor?
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/open-webui#33567