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 #11297] qwen2.5vl model crashes on small image input (height or width < 28px) #69510

New Issue
Closed
opened 2026-05-04 18:17:24 -05:00 by GiteaMirror · 2 comments
GiteaMirror commented 2026-05-04 18:17:24 -05:00
Owner
Copy Link

Originally created by @uhrinalex on GitHub (Jul 4, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/11297

What is the issue?

When using vision models like qwen2.5vl with Ollama, if an image is passed where width or height < 28px, the model runner crashes with:
panic: height:27 or width:257 must be larger than factor:28

Relevant log output

http: panic serving 127.0.0.1:57511: height:27 or width:257 must be larger than factor:28
...
panic ...
github.com/ollama/ollama/model/models/qwen25vl.(*ImageProcessor).SmartResize

OS

Windows 11

GPU

No response

CPU

Snapdragon X Elite (arm)

Ollama version

0.9.5

Originally created by @uhrinalex on GitHub (Jul 4, 2025). Original GitHub issue: https://github.com/ollama/ollama/issues/11297 ### What is the issue? When using vision models like qwen2.5vl with Ollama, if an image is passed where width or height < 28px, the model runner crashes with: `panic: height:27 or width:257 must be larger than factor:28` ### Relevant log output ```shell http: panic serving 127.0.0.1:57511: height:27 or width:257 must be larger than factor:28 ... panic ... github.com/ollama/ollama/model/models/qwen25vl.(*ImageProcessor).SmartResize ``` ### OS Windows 11 ### GPU _No response_ ### CPU Snapdragon X Elite (arm) ### Ollama version 0.9.5
GiteaMirror added the bug label 2026-05-04 18:17:24 -05:00
GiteaMirror commented 2026-05-04 18:17:29 -05:00
Author
Owner
Copy Link

@yss0729 commented on GitHub (Jul 14, 2025):

mark

<!-- gh-comment-id:3067588659 --> @yss0729 commented on GitHub (Jul 14, 2025): mark
GiteaMirror commented 2026-05-04 18:17:31 -05:00
Author
Owner
Copy Link

@mikel-brostrom commented on GitHub (Jul 15, 2025):

From the Qwen2.5vl technical report paper:

During both training and
inference, the height and width of the input images are resized to multiples of 28 before being fed into the
ViT

As a workaround you could resize the image or pad it to this size

<!-- gh-comment-id:3075556870 --> @mikel-brostrom commented on GitHub (Jul 15, 2025): From the Qwen2.5vl technical report paper: > During both training and > inference, the height and width of the input images are resized to multiples of 28 before being fed into the > ViT As a workaround you could resize the image or pad it to this size
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#69510
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?