github-starred/open-webui
2
0
Fork 0
mirror of https://github.com/open-webui/open-webui.git synced 2026-05-08 21:09:41 -05:00
Code Issues 1.3k Packages Projects Releases 126 Wiki Activity

[GH-ISSUE #16236] issue: Tasks don't work on large context prompts (Renaming chat, etc) #17835

New Issue
Closed
opened 2026-04-19 23:43:33 -05:00 by GiteaMirror · 7 comments
GiteaMirror commented 2026-04-19 23:43:33 -05:00
Owner
Copy Link

Originally created by @frenzybiscuit on GitHub (Aug 2, 2025).
Original GitHub issue: https://github.com/open-webui/open-webui/issues/16236

Check Existing Issues

  • I have searched the existing issues and discussions.
  • I am using the latest version of Open WebUI.

Installation Method

Docker

Open WebUI Version

0.6.18

Ollama Version (if applicable)

No response

Operating System

debian 12

Browser (if applicable)

No response

Confirmation

  • I have read and followed all instructions in README.md.
  • I am using the latest version of both Open WebUI and Ollama.
  • I have included the browser console logs.
  • I have included the Docker container logs.
  • I have provided every relevant configuration, setting, and environment variable used in my setup.
  • I have clearly listed every relevant configuration, custom setting, environment variable, and command-line option that influences my setup (such as Docker Compose overrides, .env values, browser settings, authentication configurations, etc).
  • I have documented step-by-step reproduction instructions that are precise, sequential, and leave nothing to interpretation. My steps:
  • Start with the initial platform/version/OS and dependencies used,
  • Specify exact install/launch/configure commands,
  • List URLs visited, user input (incl. example values/emails/passwords if needed),
  • Describe all options and toggles enabled or changed,
  • Include any files or environmental changes,
  • Identify the expected and actual result at each stage,
  • Ensure any reasonably skilled user can follow and hit the same issue.

Expected Behavior

Task Model works correctly

Actual Behavior

Task model doesn't work on large system prompts.

Steps to Reproduce

Use the attached system prompt with a model such as Qwen2.5 0.5B for local/external tasks and a 70B model for the actual API/Chat.

Title generation doesnt work (and i dont think tags do either)

Drew-Riley-v0.20.txt

Logs & Screenshots

Image

When using a smaller system prompt, it works:

Image

Additional Information

Running the following:

Qwen2.5 0.5B model for tasks with 64k context.
Animus V7.0 70B 5.25BPW for main API with tabbyAPI with 24k context

Task page:

Image
Originally created by @frenzybiscuit on GitHub (Aug 2, 2025). Original GitHub issue: https://github.com/open-webui/open-webui/issues/16236 ### Check Existing Issues - [x] I have searched the existing issues and discussions. - [x] I am using the latest version of Open WebUI. ### Installation Method Docker ### Open WebUI Version 0.6.18 ### Ollama Version (if applicable) _No response_ ### Operating System debian 12 ### Browser (if applicable) _No response_ ### Confirmation - [x] I have read and followed all instructions in `README.md`. - [x] I am using the latest version of **both** Open WebUI and Ollama. - [x] I have included the browser console logs. - [x] I have included the Docker container logs. - [x] I have **provided every relevant configuration, setting, and environment variable used in my setup.** - [x] I have clearly **listed every relevant configuration, custom setting, environment variable, and command-line option that influences my setup** (such as Docker Compose overrides, .env values, browser settings, authentication configurations, etc). - [x] I have documented **step-by-step reproduction instructions that are precise, sequential, and leave nothing to interpretation**. My steps: - Start with the initial platform/version/OS and dependencies used, - Specify exact install/launch/configure commands, - List URLs visited, user input (incl. example values/emails/passwords if needed), - Describe all options and toggles enabled or changed, - Include any files or environmental changes, - Identify the expected and actual result at each stage, - Ensure any reasonably skilled user can follow and hit the same issue. ### Expected Behavior Task Model works correctly ### Actual Behavior Task model doesn't work on large system prompts. ### Steps to Reproduce Use the attached system prompt with a model such as Qwen2.5 0.5B for local/external tasks and a 70B model for the actual API/Chat. Title generation doesnt work (and i dont think tags do either) [Drew-Riley-v0.20.txt](https://github.com/user-attachments/files/21560032/Drew-Riley-v0.20.txt) ### Logs & Screenshots <img width="382" height="108" alt="Image" src="https://github.com/user-attachments/assets/d5ba4f66-6705-42ec-927e-aa1afe08072d" /> When using a smaller system prompt, it works: <img width="382" height="108" alt="Image" src="https://github.com/user-attachments/assets/aef842ce-40f6-4aad-a9be-1988ff89caf3" /> ### Additional Information Running the following: Qwen2.5 0.5B model for tasks with 64k context. Animus V7.0 70B 5.25BPW for main API with tabbyAPI with 24k context Task page: <img width="1242" height="1550" alt="Image" src="https://github.com/user-attachments/assets/607c02fb-fed2-4382-9af7-3f2743db6974" />
GiteaMirror added the bug label 2026-04-19 23:43:33 -05:00
GiteaMirror commented 2026-04-19 23:43:35 -05:00
Author
Owner
Copy Link

@frenzybiscuit commented on GitHub (Aug 2, 2025):

if you need more info let me know and I'll try to include it.

<!-- gh-comment-id:3146667445 --> @frenzybiscuit commented on GitHub (Aug 2, 2025): if you need more info let me know and I'll try to include it.
GiteaMirror commented 2026-04-19 23:43:35 -05:00
Author
Owner
Copy Link

@rgaricano commented on GitHub (Aug 3, 2025):

did you read the models README?
...
Qwen2.5-0.5B
Context Length: Full 32,768 tokens
(https://huggingface.co/Qwen/Qwen2.5-0.5B )

<!-- gh-comment-id:3148348760 --> @rgaricano commented on GitHub (Aug 3, 2025): did you read the models README? ... Qwen2.5-0.5B Context Length: Full 32,768 tokens (https://huggingface.co/Qwen/Qwen2.5-0.5B )
GiteaMirror commented 2026-04-19 23:43:36 -05:00
Author
Owner
Copy Link

@frenzybiscuit commented on GitHub (Aug 3, 2025):

did you read the models README? ... Qwen2.5-0.5B Context Length: Full 32,768 tokens (https://huggingface.co/Qwen/Qwen2.5-0.5B )

Yes, I did. I missed that because of this:

  • Long-context Support up to 128K tokens and can generate up to 8K tokens.

I will reduce the context to 32768 and get back to you on if that fixes the problem.

<!-- gh-comment-id:3148478840 --> @frenzybiscuit commented on GitHub (Aug 3, 2025): > did you read the models README? ... Qwen2.5-0.5B Context Length: Full 32,768 tokens (https://huggingface.co/Qwen/Qwen2.5-0.5B ) Yes, I did. I missed that because of this: * Long-context Support up to 128K tokens and can generate up to 8K tokens. I will reduce the context to 32768 and get back to you on if that fixes the problem.
GiteaMirror commented 2026-04-19 23:43:36 -05:00
Author
Owner
Copy Link

@frenzybiscuit commented on GitHub (Aug 3, 2025):

I reduced it to 32768, same issue.

<!-- gh-comment-id:3148483440 --> @frenzybiscuit commented on GitHub (Aug 3, 2025): I reduced it to 32768, same issue.
GiteaMirror commented 2026-04-19 23:43:37 -05:00
Author
Owner
Copy Link

@frenzybiscuit commented on GitHub (Aug 3, 2025):

I have also tried to change the API model to "current" and it still doesn't work.

<!-- gh-comment-id:3148505652 --> @frenzybiscuit commented on GitHub (Aug 3, 2025): I have also tried to change the API model to "current" and it still doesn't work.
GiteaMirror commented 2026-04-19 23:43:37 -05:00
Author
Owner
Copy Link

@frenzybiscuit commented on GitHub (Aug 3, 2025):

Here is how it should be working on long context prompts:

https://github.com/user-attachments/assets/b90bd865-3837-4bbc-a444-7cff8cbbcf62

How it actually works on long context prompts (no tasks work(titles/tags/follow-up)) and the UI freezes when leaving/entering the chat, along with actually posting the chat.

https://github.com/user-attachments/assets/c3675082-9b25-4ebc-84b5-76fe7201cf4f

<!-- gh-comment-id:3148538120 --> @frenzybiscuit commented on GitHub (Aug 3, 2025): Here is how it should be working on long context prompts: https://github.com/user-attachments/assets/b90bd865-3837-4bbc-a444-7cff8cbbcf62 How it actually works on long context prompts (no tasks work(titles/tags/follow-up)) and the UI freezes when leaving/entering the chat, along with actually posting the chat. https://github.com/user-attachments/assets/c3675082-9b25-4ebc-84b5-76fe7201cf4f
GiteaMirror commented 2026-04-19 23:43:37 -05:00
Author
Owner
Copy Link

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

As much as Qwen2.5 0.5B is a great model, it's honestly too dumb to complete the tasks with large contexts, nothing much we can do from our end. Keep us updated!

<!-- gh-comment-id:3148542752 --> @tjbck commented on GitHub (Aug 3, 2025): As much as Qwen2.5 0.5B is a great model, it's honestly too dumb to complete the tasks with large contexts, nothing much we can do from our end. Keep us updated!
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
bug
confirmed
confirmed issue
core
documentation
enhancement
good first issue
help wanted
non-core
pull-request
Mirrored from GitHub Pull Request
 python
question
testing wanted
No Label bug
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
GiteaMirror ninjasurge
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/open-webui#17835
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?