github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-07-16 21:25:46 -05:00
Code Issues 16.9k Packages Projects Releases 122 Wiki Activity

[GH-ISSUE #11699] gpt-oss:20b failes with flash attention = ture , works if false #7743

New Issue
Closed
opened 2026-04-12 19:52:29 -05:00 by GiteaMirror · 5 comments
GiteaMirror commented 2026-04-12 19:52:29 -05:00
Owner
Copy Link

Originally created by @mohammedgomaa on GitHub (Aug 5, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/11699

What is the issue?

Error: model runner has unexpectedly stopped, this may be due to resource limitations or an internal error, check ollama server logs for details
#Environment="CUDA_VISIBLE_DEVICES=0"
#Environment="CUDA_VISIBLE_DEVICES=1"
Environment="OLLAMA_NUM_PARALLEL=1"
Environment="OLLAMA_MAX_LOADED_MODELS=1"
#Environment="OLLAMA_FLASH_ATTENTION=true"
Environment="OLLAMA_HOST=0.0.0.0:11434"
Environment="CUDA_MEMORY_FRACTION=0.95"
Environment="OLLAMA_GPU_MEMORY_FRACTION=0.95"
Environment="OLLAMA_KEEP_ALIVE=5m"
Environment="OLLAMA_NUM_THREAD=6"
Environment="OLLAMA_KV_CACHE_TYPE=q4_0"
Environment="OLLAMA_CONTEXT_LENGTH=8000"
Environment="OLLAMA_GPU_ALLOW_GROWTH=true"
Environment="OLLAMA_NEW_ENGINE=true"
Environment="OLLAMA_FLASH_ATTENTION=true"

Relevant log output

OS

No response

GPU

No response

CPU

No response

Ollama version

No response

Originally created by @mohammedgomaa on GitHub (Aug 5, 2025). Original GitHub issue: https://github.com/ollama/ollama/issues/11699 ### What is the issue? Error: model runner has unexpectedly stopped, this may be due to resource limitations or an internal error, check ollama server logs for details #Environment="CUDA_VISIBLE_DEVICES=0" #Environment="CUDA_VISIBLE_DEVICES=1" Environment="OLLAMA_NUM_PARALLEL=1" Environment="OLLAMA_MAX_LOADED_MODELS=1" #Environment="OLLAMA_FLASH_ATTENTION=true" Environment="OLLAMA_HOST=0.0.0.0:11434" Environment="CUDA_MEMORY_FRACTION=0.95" Environment="OLLAMA_GPU_MEMORY_FRACTION=0.95" Environment="OLLAMA_KEEP_ALIVE=5m" Environment="OLLAMA_NUM_THREAD=6" Environment="OLLAMA_KV_CACHE_TYPE=q4_0" Environment="OLLAMA_CONTEXT_LENGTH=8000" Environment="OLLAMA_GPU_ALLOW_GROWTH=true" Environment="OLLAMA_NEW_ENGINE=true" Environment="OLLAMA_FLASH_ATTENTION=true" ### Relevant log output ```shell ``` ### OS _No response_ ### GPU _No response_ ### CPU _No response_ ### Ollama version _No response_
GiteaMirror added the bug label 2026-04-12 19:52:29 -05:00
GiteaMirror commented 2026-04-12 19:52:30 -05:00
Author
Owner
Copy Link

@bezo97 commented on GitHub (Aug 5, 2025):

I can confirm this same behavior 👍

<!-- gh-comment-id:3156730642 --> @bezo97 commented on GitHub (Aug 5, 2025): I can confirm this same behavior 👍
GiteaMirror commented 2026-04-12 19:52:31 -05:00
Author
Owner
Copy Link

@ibbobud commented on GitHub (Aug 5, 2025):

same here, but with V100 32gb... and if context pushes vram size anywhere close to 50% of vram ussage it will just go 100% cpu. go look in the discord general chat... I am aware that v100 dont support flash attention, but on any models that does not use the new ollama engine, flash attention works fine (example newest qwen models).

<!-- gh-comment-id:3156836409 --> @ibbobud commented on GitHub (Aug 5, 2025): same here, but with V100 32gb... and if context pushes vram size anywhere close to 50% of vram ussage it will just go 100% cpu. go look in the discord general chat... I am aware that v100 dont support flash attention, but on any models that does not use the new ollama engine, flash attention works fine (example newest qwen models).
GiteaMirror commented 2026-04-12 19:52:31 -05:00
Author
Owner
Copy Link

@digitalextremist commented on GitHub (Aug 6, 2025):

Not seeing version mentioned, but there was a 0.11.0 crash like this for me; then 0.11.2 ran, but used 100% CPU

For reference, my docker-compose.yml is using:

environment:
  - OLLAMA_FLASH_ATTENTION=1
  - OLLAMA_KV_CACHE_TYPE=q8_0
<!-- gh-comment-id:3157185829 --> @digitalextremist commented on GitHub (Aug 6, 2025): Not seeing version mentioned, but there was a `0.11.0` crash like this for me; then `0.11.2` ran, but used 100% CPU For reference, my `docker-compose.yml` is using: ``` environment: - OLLAMA_FLASH_ATTENTION=1 - OLLAMA_KV_CACHE_TYPE=q8_0 ```
GiteaMirror commented 2026-04-12 19:52:31 -05:00
Author
Owner
Copy Link

@thedaveCA commented on GitHub (Aug 6, 2025):

As a datapoint: 0.11.0 ran gpt-oss:20b on CPU for me, 0.11.2 on GPU. 24GB VRAM, reporting 14.8GiB in use.

OLLAMA_FLASH_ATTENTION=1

<!-- gh-comment-id:3157870248 --> @thedaveCA commented on GitHub (Aug 6, 2025): As a datapoint: `0.11.0` ran gpt-oss:20b on CPU for me, `0.11.2` on GPU. 24GB VRAM, reporting 14.8GiB in use. `OLLAMA_FLASH_ATTENTION=1`
GiteaMirror commented 2026-04-12 19:52:32 -05:00
Author
Owner
Copy Link

@jessegross commented on GitHub (Aug 6, 2025):

Fixed in 0.11.2

<!-- gh-comment-id:3161682790 --> @jessegross commented on GitHub (Aug 6, 2025): Fixed in 0.11.2
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#7743
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?