github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-07 00:22:43 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #11828] Qwen3-4B-Instruct-2507-GGUF:UD-Q4_K_XL - talking crazy #7849

New Issue
Closed
opened 2026-04-12 20:00:50 -05:00 by GiteaMirror · 8 comments
GiteaMirror commented 2026-04-12 20:00:50 -05:00
Owner
Copy Link

Originally created by @slepkaviba on GitHub (Aug 9, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/11828

What is the issue?

As mentioned hf.co/unsloth/Qwen3-4B-Instruct-2507-GGUF:UD-Q4_K_XL starts to talk about random things with any prompt. But qwen3:4b-instruct-2507-q4_K_M seems to be working fine.

Is that more issue with the model, or the support of it?

Image

Relevant log output

OS

Windows/Linux

GPU

RTX 3080 / RTX 4060 Ti

CPU

No response

Ollama version

0.11.4

Originally created by @slepkaviba on GitHub (Aug 9, 2025). Original GitHub issue: https://github.com/ollama/ollama/issues/11828 ### What is the issue? As mentioned hf.co/unsloth/Qwen3-4B-Instruct-2507-GGUF:UD-Q4_K_XL starts to talk about random things with any prompt. But qwen3:4b-instruct-2507-q4_K_M seems to be working fine. Is that more issue with the model, or the support of it? <img width="1174" height="178" alt="Image" src="https://github.com/user-attachments/assets/d5c02f92-7068-4ec8-831e-0015ea1b6d37" /> ### Relevant log output ```shell ``` ### OS Windows/Linux ### GPU RTX 3080 / RTX 4060 Ti ### CPU _No response_ ### Ollama version 0.11.4
GiteaMirror added the bug label 2026-04-12 20:00:50 -05:00
GiteaMirror commented 2026-04-12 20:00:51 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Aug 9, 2025):

Add a template.

<!-- gh-comment-id:3170477716 --> @rick-github commented on GitHub (Aug 9, 2025): Add a template.
GiteaMirror commented 2026-04-12 20:00:52 -05:00
Author
Owner
Copy Link

@slepkaviba commented on GitHub (Aug 9, 2025):

I am little confused, as Unsloth gives example:
<|im_start|>user Hey there!<|im_end|> <|im_start|>assistant What is 1+1?<|im_end|> <|im_start|>user 2<|im_end|> <|im_start|>assistant
But in Ollama's docs templates expect parameters... Is there a way to learn how to translate a to b?

There is ChatML example:
{{- range .Messages }}<|im_start|>{{ .Role }} {{ .Content }}<|im_end|> {{ end }}<|im_start|>assistant

But it seems little different..

<!-- gh-comment-id:3170495667 --> @slepkaviba commented on GitHub (Aug 9, 2025): I am little confused, as Unsloth gives example: ` <|im_start|>user Hey there!<|im_end|> <|im_start|>assistant What is 1+1?<|im_end|> <|im_start|>user 2<|im_end|> <|im_start|>assistant ` But [in Ollama's docs](https://github.com/ollama/ollama/blob/main/docs/template.md ) templates expect parameters... Is there a way to learn how to translate a to b? There is ChatML example: ` {{- range .Messages }}<|im_start|>{{ .Role }} {{ .Content }}<|im_end|> {{ end }}<|im_start|>assistant ` But it seems little different..
GiteaMirror commented 2026-04-12 20:00:52 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Aug 9, 2025):

Take the template from qwen3:4b-instruct-2507-q4_K_M and add it to your imported model.

<!-- gh-comment-id:3170496679 --> @rick-github commented on GitHub (Aug 9, 2025): Take the template from qwen3:4b-instruct-2507-q4_K_M and add it to your imported model.
GiteaMirror commented 2026-04-12 20:00:53 -05:00
Author
Owner
Copy Link

@slepkaviba commented on GitHub (Aug 9, 2025):

Ok... And in the "long term" is it something that would be fixed on the model?

<!-- gh-comment-id:3170499266 --> @slepkaviba commented on GitHub (Aug 9, 2025): Ok... And in the "long term" is it something that would be fixed on the model?
GiteaMirror commented 2026-04-12 20:00:54 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Aug 9, 2025):

You would have to ask unsloth.

<!-- gh-comment-id:3170500046 --> @rick-github commented on GitHub (Aug 9, 2025): You would have to ask unsloth.
GiteaMirror commented 2026-04-12 20:00:54 -05:00
Author
Owner
Copy Link

@shimmyshimmer commented on GitHub (Aug 20, 2025):

We added the chat template now @slepkaviba so should be fixed

<!-- gh-comment-id:3204355184 --> @shimmyshimmer commented on GitHub (Aug 20, 2025): We added the chat template now @slepkaviba so should be fixed
GiteaMirror commented 2026-04-12 20:00:55 -05:00
Author
Owner
Copy Link

@slepkaviba commented on GitHub (Aug 20, 2025):

You are amazing! Thank you!

<!-- gh-comment-id:3204366121 --> @slepkaviba commented on GitHub (Aug 20, 2025): You are amazing! Thank you!
GiteaMirror commented 2026-04-12 20:00:56 -05:00
Author
Owner
Copy Link

@slepkaviba commented on GitHub (Aug 20, 2025):

Will same template be added to the Thinking variant?

<!-- gh-comment-id:3204369576 --> @slepkaviba commented on GitHub (Aug 20, 2025): Will same template be added to the Thinking variant?
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
amd
api
app
bug
build
cli
cloud
compatibility
context-length
create
docker
documentation
embeddings
feature request
feedback wanted
good first issue
gpt-oss
gpu
harmony
help wanted
image
install
intel
js
launch
linux
macos
memory
mlx
model
needs more info
networking
nvidia
ollama.com
performance
pull-request
Mirrored from GitHub Pull Request
 python
question
registry
rendering
thinking
tools
top
vulkan
windows
wsl
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/ollama#7849
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?