github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-07-15 20:54:54 -05:00
Code Issues 16.9k Packages Projects Releases 122 Wiki Activity

[GH-ISSUE #8620] Add support fo Qwen 2.5 VL models (3B, 7B and 32B) instruct versions #83264

New Issue
Closed
opened 2026-05-09 17:37:57 -05:00 by GiteaMirror · 29 comments
GiteaMirror commented 2026-05-09 17:37:57 -05:00
Owner
Copy Link

Originally created by @YarvixPA on GitHub (Jan 27, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/8620

Hello, now hours after being released, I would like to suggest that you add support for the Qwen2.5-VL models.

Qwen2.5-VL - Hugginface collection

Originally created by @YarvixPA on GitHub (Jan 27, 2025). Original GitHub issue: https://github.com/ollama/ollama/issues/8620 Hello, now hours after being released, I would like to suggest that you add support for the Qwen2.5-VL models. **[Qwen2.5-VL - Hugginface collection](https://huggingface.co/collections/Qwen/qwen25-vl-6795ffac22b334a837c0f9a5)**
GiteaMirror added the model label 2026-05-09 17:37:57 -05:00
GiteaMirror commented 2026-05-09 17:37:59 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Jan 27, 2025):

https://github.com/ollama/ollama/issues/6564

<!-- gh-comment-id:2617026007 --> @rick-github commented on GitHub (Jan 27, 2025): https://github.com/ollama/ollama/issues/6564
GiteaMirror commented 2026-05-09 17:38:01 -05:00
Author
Owner
Copy Link

@YarvixPA commented on GitHub (Jan 27, 2025):

It is not the same model, that VL is from Qwen 2.0. The one I am suggesting is Qwen2.5

<!-- gh-comment-id:2617029336 --> @YarvixPA commented on GitHub (Jan 27, 2025): It is not the same model, that VL is from Qwen 2.0. The one I am suggesting is Qwen2.5
GiteaMirror commented 2026-05-09 17:38:04 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Jan 27, 2025):

The effort to support it is the same. When support for qwen2-vl projector is done, support for qwen2.5-vl projector is incremental.

<!-- gh-comment-id:2617032004 --> @rick-github commented on GitHub (Jan 27, 2025): The effort to support it is the same. When support for qwen2-vl projector is done, support for qwen2.5-vl projector is incremental.
GiteaMirror commented 2026-05-09 17:38:06 -05:00
Author
Owner
Copy Link

@3unnycheung commented on GitHub (Jan 28, 2025):

8c9fb8eb73

<!-- gh-comment-id:2618904260 --> @3unnycheung commented on GitHub (Jan 28, 2025): https://github.com/ollama/ollama/commit/8c9fb8eb73afc220e8bf99772572096b6498b748
GiteaMirror commented 2026-05-09 17:38:09 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Jan 28, 2025):

8c9fb8e is refactoring, actual support is still pending in https://github.com/ollama/ollama/pull/8113

<!-- gh-comment-id:2618916642 --> @rick-github commented on GitHub (Jan 28, 2025): 8c9fb8e is refactoring, actual support is still pending in https://github.com/ollama/ollama/pull/8113
GiteaMirror commented 2026-05-09 17:38:12 -05:00
Author
Owner
Copy Link

@3unnycheung commented on GitHub (Jan 28, 2025):

thx, i'm looking forward to it

<!-- gh-comment-id:2618927098 --> @3unnycheung commented on GitHub (Jan 28, 2025): thx, i'm looking forward to it
GiteaMirror commented 2026-05-09 17:38:12 -05:00
Author
Owner
Copy Link

@zytoh0 commented on GitHub (Jan 30, 2025):

Hello, now hours after being released, I would like to suggest that you add support for the Qwen2.5-VL models.

Qwen2.5-VL - Hugginface collection

There're only 3B, 7B and 72B instruct versions? Where's the 32B instruct version? Please support them all :)

Image

<!-- gh-comment-id:2623597309 --> @zytoh0 commented on GitHub (Jan 30, 2025): > Hello, now hours after being released, I would like to suggest that you add support for the Qwen2.5-VL models. > > **[Qwen2.5-VL - Hugginface collection](https://huggingface.co/collections/Qwen/qwen25-vl-6795ffac22b334a837c0f9a5)** There're only 3B, 7B and 72B instruct versions? Where's the 32B instruct version? Please support them all :) ![Image](https://github.com/user-attachments/assets/e869bb10-8687-47f4-aa4a-a182af1fa98a)
GiteaMirror commented 2026-05-09 17:38:13 -05:00
Author
Owner
Copy Link

@seblessa commented on GitHub (Jan 30, 2025):

Hello, do we know when the Qwen2.5-VL-Instruct family is getting released too Ollama ?

<!-- gh-comment-id:2624257634 --> @seblessa commented on GitHub (Jan 30, 2025): Hello, do we know when the Qwen2.5-VL-Instruct family is getting released too Ollama ?
GiteaMirror commented 2026-05-09 17:38:13 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Jan 30, 2025):

After the code is merged.

<!-- gh-comment-id:2624259973 --> @rick-github commented on GitHub (Jan 30, 2025): After the code is merged.
GiteaMirror commented 2026-05-09 17:38:13 -05:00
Author
Owner
Copy Link

@MoRocety commented on GitHub (Jan 30, 2025):

has the pull request been made? I am assuming its close to being merged?

<!-- gh-comment-id:2624949754 --> @MoRocety commented on GitHub (Jan 30, 2025): has the pull request been made? I am assuming its close to being merged?
GiteaMirror commented 2026-05-09 17:38:14 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Jan 30, 2025):

The PR is still in draft mode.

<!-- gh-comment-id:2624954832 --> @rick-github commented on GitHub (Jan 30, 2025): The PR is still in draft mode.
GiteaMirror commented 2026-05-09 17:38:14 -05:00
Author
Owner
Copy Link

@MoRocety commented on GitHub (Jan 30, 2025):

when's it expected to be merged?

<!-- gh-comment-id:2624977878 --> @MoRocety commented on GitHub (Jan 30, 2025): when's it expected to be merged?
GiteaMirror commented 2026-05-09 17:38:15 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Jan 30, 2025):

After it's approved.

<!-- gh-comment-id:2624982940 --> @rick-github commented on GitHub (Jan 30, 2025): After it's approved.
GiteaMirror commented 2026-05-09 17:38:15 -05:00
Author
Owner
Copy Link

@nguyenhongson1902 commented on GitHub (Jan 30, 2025):

I leave a comment here to get notified when there's support for this model with ollama. Can't wait to give it a try :)

<!-- gh-comment-id:2625090393 --> @nguyenhongson1902 commented on GitHub (Jan 30, 2025): I leave a comment here to get notified when there's support for this model with ollama. Can't wait to give it a try :)
GiteaMirror commented 2026-05-09 17:38:16 -05:00
Author
Owner
Copy Link

@pinballelectronica commented on GitHub (Jan 30, 2025):

After you approve the PR when will my pizza be ready?

<!-- gh-comment-id:2625103584 --> @pinballelectronica commented on GitHub (Jan 30, 2025): After you approve the PR when will my pizza be ready?
GiteaMirror commented 2026-05-09 17:38:16 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Jan 30, 2025):

We'll stick the pizza base in an oven surrounded by GPUs loading qwen2.5-vl, it will be flash cooked in 0.0001 seconds.

<!-- gh-comment-id:2625109087 --> @rick-github commented on GitHub (Jan 30, 2025): We'll stick the pizza base in an oven surrounded by GPUs loading qwen2.5-vl, it will be flash cooked in 0.0001 seconds.
GiteaMirror commented 2026-05-09 17:38:17 -05:00
Author
Owner
Copy Link

@nguyenhongson1902 commented on GitHub (Jan 30, 2025):

We'll stick the pizza base in an oven surrounded by GPUs loading qwen2.5-vl, it will be flash cooked in 0.0001 seconds.

You have a great sense of humor :)

<!-- gh-comment-id:2625114605 --> @nguyenhongson1902 commented on GitHub (Jan 30, 2025): > We'll stick the pizza base in an oven surrounded by GPUs loading qwen2.5-vl, it will be flash cooked in 0.0001 seconds. You have a great sense of humor :)
GiteaMirror commented 2026-05-09 17:38:17 -05:00
Author
Owner
Copy Link

@pdevine commented on GitHub (Jan 30, 2025):

TLDR; we're working on this, it's going to take a while

Long version: we tried getting this working w/ llama.cpp but it was very hacked together and we couldn't get reasonable quality out of it. We could have shipped that version but I don't think anyone would have been particularly happy. We are looking to get it working w/ the new engine, but we're still a ways away from getting the new engine out (I could give a timeline for that, but it's invariably going to be wrong).

<!-- gh-comment-id:2625327610 --> @pdevine commented on GitHub (Jan 30, 2025): TLDR; we're working on this, it's going to take a while Long version: we tried getting this working w/ llama.cpp but it was very hacked together and we couldn't get reasonable quality out of it. We could have shipped that version but I don't think anyone would have been particularly happy. We _are_ looking to get it working w/ the new engine, but we're still a ways away from getting the new engine out (I could give a timeline for that, but it's invariably going to be wrong).
GiteaMirror commented 2026-05-09 17:38:19 -05:00
Author
Owner
Copy Link

@pdevine commented on GitHub (Jan 30, 2025):

Also, I'm going to update #6564 to reflect 2.5 instead of 2.0. I don't think we need two issues.

<!-- gh-comment-id:2625350951 --> @pdevine commented on GitHub (Jan 30, 2025): Also, I'm going to update #6564 to reflect 2.5 instead of 2.0. I don't think we need two issues.
GiteaMirror commented 2026-05-09 17:38:20 -05:00
Author
Owner
Copy Link

@wenshyan commented on GitHub (Feb 25, 2025):

Hello All.

I would like to ask if there are any updates to the addition of QWEN 2.5 VL to the Ollama Library ?

<!-- gh-comment-id:2682602764 --> @wenshyan commented on GitHub (Feb 25, 2025): Hello All. I would like to ask if there are any updates to the addition of QWEN 2.5 VL to the Ollama Library ?
GiteaMirror commented 2026-05-09 17:38:21 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Feb 25, 2025):

https://github.com/ollama/ollama/issues/6564#issuecomment-2660054413

<!-- gh-comment-id:2682626199 --> @rick-github commented on GitHub (Feb 25, 2025): https://github.com/ollama/ollama/issues/6564#issuecomment-2660054413
GiteaMirror commented 2026-05-09 17:38:21 -05:00
Author
Owner
Copy Link

@wenshyan commented on GitHub (Feb 25, 2025):

Thank you for the quick response. Looking forward to have the model in ollama. We are currently testing this on a NVIDIA Jetson Orin Nano.

<!-- gh-comment-id:2682672303 --> @wenshyan commented on GitHub (Feb 25, 2025): Thank you for the quick response. Looking forward to have the model in ollama. We are currently testing this on a NVIDIA Jetson Orin Nano.
GiteaMirror commented 2026-05-09 17:38:22 -05:00
Author
Owner
Copy Link

@pvtoan commented on GitHub (Mar 6, 2025):

Hi @wenshyan ,

Can I know if you have finished testing Qwen 2.5 VL on Jetson Orin Nano yet?

What exact models (3B, 7B, ...; original or quantized models) you are able to run on Jetson and how about testing results?

<!-- gh-comment-id:2703188885 --> @pvtoan commented on GitHub (Mar 6, 2025): Hi @wenshyan , Can I know if you have finished testing Qwen 2.5 VL on Jetson Orin Nano yet? What exact models (3B, 7B, ...; original or quantized models) you are able to run on Jetson and how about testing results?
GiteaMirror commented 2026-05-09 17:38:23 -05:00
Author
Owner
Copy Link

@rogerc66 commented on GitHub (Mar 25, 2025):

mark to get notice.

<!-- gh-comment-id:2751711023 --> @rogerc66 commented on GitHub (Mar 25, 2025): mark to get notice.
GiteaMirror commented 2026-05-09 17:38:23 -05:00
Author
Owner
Copy Link

@imadcat commented on GitHub (Apr 13, 2025):

any progress?

<!-- gh-comment-id:2799948854 --> @imadcat commented on GitHub (Apr 13, 2025): any progress?
GiteaMirror commented 2026-05-09 17:38:24 -05:00
Author
Owner
Copy Link

@10067 commented on GitHub (Sep 17, 2025):

mark

<!-- gh-comment-id:3301427828 --> @10067 commented on GitHub (Sep 17, 2025): mark
GiteaMirror commented 2026-05-09 17:38:25 -05:00
Author
Owner
Copy Link

@pinballelectronica commented on GitHub (Sep 21, 2025):

There's nothing magical about Ollama. You don't need to wait for someone to do it for you. You create a manifest and point them to gguf weights. It's really easy to take any one of those models and make them Ollama compatible. Plus you can choose your own quant size and temp, context size etc.

_FROM ./model.gguf

PARAMETER num_ctx 4096
PARAMETER temperature 0.7

TEMPLATE """
{{ if .System }}System: {{ .System }} {{ end }}
{{ if .Prompt }}User: {{ .Prompt }} {{ end }}
Assistant:
"""_

<!-- gh-comment-id:3315923380 --> @pinballelectronica commented on GitHub (Sep 21, 2025): There's nothing magical about Ollama. You don't need to wait for someone to do it for you. You create a manifest and point them to gguf weights. It's really easy to take any one of those models and make them Ollama compatible. Plus you can choose your own quant size and temp, context size etc. _FROM ./model.gguf PARAMETER num_ctx 4096 PARAMETER temperature 0.7 TEMPLATE """ {{ if .System }}System: {{ .System }} {{ end }} {{ if .Prompt }}User: {{ .Prompt }} {{ end }} Assistant: """_
GiteaMirror commented 2026-05-09 17:38:25 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Sep 21, 2025):

These models are already supported, there's no need to import.

https://ollama.com/library/qwen2.5vl

<!-- gh-comment-id:3315944976 --> @rick-github commented on GitHub (Sep 21, 2025): These models are already supported, there's no need to import. https://ollama.com/library/qwen2.5vl
GiteaMirror commented 2026-05-09 17:38:28 -05:00
Author
Owner
Copy Link

@pinballelectronica commented on GitHub (Sep 21, 2025):

"¯_(ツ)_/¯"

<!-- gh-comment-id:3315945996 --> @pinballelectronica commented on GitHub (Sep 21, 2025): "¯\_(ツ)_/¯"
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 model
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#83264
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?