github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-07 08:30:05 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #9728] Gemma3 wrong context length #52871

New Issue
Closed
opened 2026-04-29 01:14:35 -05:00 by GiteaMirror · 3 comments
GiteaMirror commented 2026-04-29 01:14:35 -05:00
Owner
Copy Link

Originally created by @wizardbc on GitHub (Mar 13, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/9728

What is the issue?

Gemma3 4b, 12b, 27b models' gemma3.context_length is NOT 8192.
It should be 131072.

Relevant log output

OS

No response

GPU

No response

CPU

No response

Ollama version

No response

Originally created by @wizardbc on GitHub (Mar 13, 2025). Original GitHub issue: https://github.com/ollama/ollama/issues/9728 ### What is the issue? Gemma3 4b, 12b, 27b models' `gemma3.context_length` is NOT `8192`. It should be `131072`. ### Relevant log output ```shell ``` ### OS _No response_ ### GPU _No response_ ### CPU _No response_ ### Ollama version _No response_
GiteaMirror added the bug label 2026-04-29 01:14:35 -05:00
GiteaMirror commented 2026-04-29 01:14:38 -05:00
Author
Owner
Copy Link

@nistvan86 commented on GitHub (Mar 13, 2025):

I think this is the default amount it will allocate and you can increase it. But this context length requires a lot of additional VRAM, I think I read somewhere 20+GB VRAM for 128k.

Also it seems Ollama is not really efficient with handling kv caches just yet and needs optimizations. from Reddit

<!-- gh-comment-id:2721252279 --> @nistvan86 commented on GitHub (Mar 13, 2025): I think this is the default amount it will allocate and you can increase it. But this context length requires a lot of additional VRAM, I think I read somewhere 20+GB VRAM for 128k. Also it seems Ollama is not really efficient with handling kv caches just yet and needs optimizations. [from Reddit](https://www.reddit.com/r/ollama/comments/1j9kv29/comment/mhfe6ok/?utm_source=share&utm_medium=web3x&utm_name=web3xcss&utm_term=1&utm_content=share_button)
GiteaMirror commented 2026-04-29 01:14:39 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Mar 13, 2025):

#9702

<!-- gh-comment-id:2721416473 --> @rick-github commented on GitHub (Mar 13, 2025): #9702
GiteaMirror commented 2026-04-29 01:14:39 -05:00
Author
Owner
Copy Link

@pdevine commented on GitHub (Mar 13, 2025):

See my other comment in #9702... hopefully clarifies that yes, you can set it to 128k right now.

<!-- gh-comment-id:2722846915 --> @pdevine commented on GitHub (Mar 13, 2025): See my other comment in #9702... hopefully clarifies that yes, you can set it to 128k right now.
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#52871
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?