github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-24 14:43:39 -05:00
Code Issues 16.9k Packages Projects Releases 122 Wiki Activity

[GH-ISSUE #8335] Make flash attention configurable via UI or enable by default #5342

New Issue
Closed
opened 2026-04-12 16:32:32 -05:00 by GiteaMirror · 3 comments
GiteaMirror commented 2026-04-12 16:32:32 -05:00
Owner
Copy Link

Originally created by @HDembinski on GitHub (Jan 7, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/8335

Hi, I love Ollama, excellent work. It makes using LLMs really beginner friendly, but does impose any limits on power usage.

I recently learned about flash attention and found out from reading the FAQ that Ollama supports this. As flash attention is important to support large contexts and can speed up models considerably, it would be great if the option to enable flash attention would be more easily accessible.

I am on Windows, and the Ollama Server has a small icon in the notification area. It would be great if you could add a checkbox to enable flash attention and set the KV cache quantization there.

Originally created by @HDembinski on GitHub (Jan 7, 2025). Original GitHub issue: https://github.com/ollama/ollama/issues/8335 Hi, I love Ollama, excellent work. It makes using LLMs really beginner friendly, but does impose any limits on power usage. I recently learned about flash attention and found out from reading the FAQ that Ollama supports this. As flash attention is important to support large contexts and can speed up models considerably, it would be great if the option to enable flash attention would be more easily accessible. I am on Windows, and the Ollama Server has a small icon in the notification area. It would be great if you could add a checkbox to enable flash attention and set the KV cache quantization there.
GiteaMirror added the feature request label 2026-04-12 16:32:32 -05:00
GiteaMirror commented 2026-04-12 16:32:33 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Feb 9, 2026):

This is the default for most of the newer models.

<!-- gh-comment-id:3872816223 --> @rick-github commented on GitHub (Feb 9, 2026): This is the [default](https://github.com/ollama/ollama/pulls?q=is%3Apr+flash+attention+default+is%3Aclosed) for most of the newer models.
GiteaMirror commented 2026-04-12 16:32:34 -05:00
Author
Owner
Copy Link

@zoltanf commented on GitHub (Apr 11, 2026):

Not for gemma4 unfortunately, and when it's enabled trough env. variable it gives a 50% tokens/sec boost on my M1Pro mac.
Also super hard to enable when ollama is running trough brew service.

<!-- gh-comment-id:4229775693 --> @zoltanf commented on GitHub (Apr 11, 2026): Not for gemma4 unfortunately, and when it's enabled trough env. variable it gives a 50% tokens/sec boost on my M1Pro mac. Also super hard to enable when ollama is running trough brew service.
GiteaMirror commented 2026-04-12 16:32:34 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Apr 11, 2026):

https://github.com/ollama/ollama/pull/15378

<!-- gh-comment-id:4229776888 --> @rick-github commented on GitHub (Apr 11, 2026): https://github.com/ollama/ollama/pull/15378
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 feature request
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#5342
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?