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 #6255] Update LLaVA to LLaVA OneVision #65952

New Issue
Open
opened 2026-05-03 23:18:34 -05:00 by GiteaMirror · 4 comments
GiteaMirror commented 2026-05-03 23:18:34 -05:00
Owner
Copy Link

Originally created by @alexrah on GitHub (Aug 8, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/6255

LLaVA has a new version called OneVision which was released 2024/08/06

HuggingFace
GitHub
Release Notes

key features:

  • Supports various input resolutions up to 2304 * 2304 pixels.
  • Single image input is represented by 729 * (9+1) tokens at most under anyres_max_9 mode.
  • Supports multi-image and video inputs. Multi-image input is represented by 729 token for each image, and video input is represented by 196 token for each frame.
  • Available in three sizes: 0.5B, 7B and 72B parameter versions, fit for different memory and inference latency requirements.
  • better support for Set-of-mark prompting
  • and more...
Originally created by @alexrah on GitHub (Aug 8, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/6255 LLaVA has a new version called OneVision which was released 2024/08/06 [HuggingFace](https://huggingface.co/collections/lmms-lab/llava-onevision-66a259c3526e15166d6bba37) [GitHub](https://github.com/LLaVA-VL/LLaVA-NeXT/) [Release Notes](https://llava-vl.github.io/blog/2024-08-05-llava-onevision/) ### key features: - Supports various input resolutions up to 2304 * 2304 pixels. - Single image input is represented by 729 * (9+1) tokens at most under anyres_max_9 mode. - Supports multi-image and video inputs. Multi-image input is represented by 729 token for each image, and video input is represented by 196 token for each frame. - Available in three sizes: 0.5B, 7B and 72B parameter versions, fit for different memory and inference latency requirements. - better support for Set-of-mark prompting - and more...
GiteaMirror added the model label 2026-05-03 23:18:34 -05:00
GiteaMirror commented 2026-05-03 23:18:39 -05:00
Author
Owner
Copy Link

@evolu8 commented on GitHub (Aug 31, 2024):

a Q4 of the Qwen 72b version is a world apart from all the other open models I've tested and yet is workable for small business / many (not all) private budgets re hardware. Fitting on two large or one very large GPU and running at decent speed. It might not be getting the press, but this is the game changer for many. Please let us know if there are challenges the community can help with.

<!-- gh-comment-id:2322825874 --> @evolu8 commented on GitHub (Aug 31, 2024): a Q4 of the Qwen 72b version is a world apart from all the other open models I've tested and yet is workable for small business / many (not all) private budgets re hardware. Fitting on two large or one very large GPU and running at decent speed. It might not be getting the press, but this is the game changer for many. Please let us know if there are challenges the community can help with.
GiteaMirror commented 2026-05-03 23:18:41 -05:00
Author
Owner
Copy Link

@ChieF-TroN commented on GitHub (Sep 5, 2024):

Showing my support for adding support for LLaVa-OneVision. This is leagues better than any previous LLaVa model.

<!-- gh-comment-id:2330650314 --> @ChieF-TroN commented on GitHub (Sep 5, 2024): Showing my support for adding support for LLaVa-OneVision. This is leagues better than any previous LLaVa model.
GiteaMirror commented 2026-05-03 23:18:42 -05:00
Author
Owner
Copy Link

@jignnsd commented on GitHub (Sep 5, 2024):

The vision models in ollama have very limited capabilities. Will be very useful to add better vision models like this one, so we keep up with the latest developments.
Thanks

<!-- gh-comment-id:2331733417 --> @jignnsd commented on GitHub (Sep 5, 2024): The vision models in ollama have very limited capabilities. Will be very useful to add better vision models like this one, so we keep up with the latest developments. Thanks
GiteaMirror commented 2026-05-03 23:18:43 -05:00
Author
Owner
Copy Link

@blacklig commented on GitHub (Sep 12, 2024):

yes please.. at least 7B SI

<!-- gh-comment-id:2347352425 --> @blacklig commented on GitHub (Sep 12, 2024): yes please.. at least 7B SI
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#65952
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?