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 #4783] How do I customize the number of layers to be loaded on GPU? #3012

New Issue
Closed
opened 2026-04-12 13:25:02 -05:00 by GiteaMirror · 1 comment
GiteaMirror commented 2026-04-12 13:25:02 -05:00
Owner
Copy Link

Originally created by @lingyezhixing on GitHub (Jun 2, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/4783

Originally assigned to: @dhiltgen on GitHub.

I tried adding num_gpu parameters to the modelfile, but it didn't seem to work.
My graphics memory still has 1.4GB left, but I can't make use of it.
What makes me even more annoyed is that only one layer is placed on the CPU, but this greatly reduces the inference speed.

Originally created by @lingyezhixing on GitHub (Jun 2, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/4783 Originally assigned to: @dhiltgen on GitHub. I tried adding num_gpu parameters to the modelfile, but it didn't seem to work. My graphics memory still has 1.4GB left, but I can't make use of it. What makes me even more annoyed is that only one layer is placed on the CPU, but this greatly reduces the inference speed.
GiteaMirror added the questionneeds more info labels 2026-04-12 13:25:02 -05:00
GiteaMirror commented 2026-04-12 13:25:07 -05:00
Author
Owner
Copy Link

@dhiltgen commented on GitHub (Jun 18, 2024):

Can you try setting num_gpu in the API request?

curl http://localhost:11434/api/generate -d '{
  "model": "myCustomModel",
  "prompt": "Why is the sky blue?",
  "stream": false, "options": {"num_gpu": 5}
}'

Then check ollama ps and the server log to see if it worked. If the system isn't behaving properly, please share the logs and the repro scenario so we can try to understand what's going wrong. If that does work, but the modelfile setting isn't working, please share a minimal modelfile that repro's the problem.

<!-- gh-comment-id:2177021629 --> @dhiltgen commented on GitHub (Jun 18, 2024): Can you try setting num_gpu in the API request? ``` curl http://localhost:11434/api/generate -d '{ "model": "myCustomModel", "prompt": "Why is the sky blue?", "stream": false, "options": {"num_gpu": 5} }' ``` Then check `ollama ps` and the server log to see if it worked. If the system isn't behaving properly, please share the logs and the repro scenario so we can try to understand what's going wrong. If that does work, but the modelfile setting isn't working, please share a minimal modelfile that repro's the problem.
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 needs more info question
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#3012
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?