github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-07-17 21:55:43 -05:00
Code Issues 16.9k Packages Projects Releases 122 Wiki Activity

[PR #10700] [MERGED] Estimate projector memory correctly for Ollama engine #23872

New Issue
Closed
opened 2026-04-19 17:15:58 -05:00 by GiteaMirror · 0 comments
GiteaMirror commented 2026-04-19 17:15:58 -05:00
Owner
Copy Link

📋 Pull Request Information

Original PR: https://github.com/ollama/ollama/pull/10700
Author: @jessegross
Created: 5/13/2025
Status: ✅ Merged
Merged: 5/19/2025
Merged by: @jessegross

Base: main ← Head: jessegross/vision_graph

📝 Commits (2)

  • bd4887c llm: Consistently track unassigned model data
  • d3a7ce2 llm: Estimate projector memory correctly for Ollama engine

📊 Changes

1 file changed (+28 additions, -20 deletions)

View changed files

📝 llm/memory.go (+28 -20)

📄 Description

The Llama engine always places vision projectors on the first GPU if one exists. However, the Ollama engine groups it with the output layer, which means the projector is only offloaded if all other layers are offloaded. The memory estimation code always assumes the former layout - this changes it to use the correct layout based on the engine.

This addresses two impacts of the current behavior:

  • In multi-GPU setups, we can crash with OOM errors when we try to allocate memory on a full GPU while another still has space.
  • If the vision projector is large, it may prevent us from offloading anything when we could have fit some of the text layers.

🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.

## 📋 Pull Request Information **Original PR:** https://github.com/ollama/ollama/pull/10700 **Author:** [@jessegross](https://github.com/jessegross) **Created:** 5/13/2025 **Status:** ✅ Merged **Merged:** 5/19/2025 **Merged by:** [@jessegross](https://github.com/jessegross) **Base:** `main` ← **Head:** `jessegross/vision_graph` --- ### 📝 Commits (2) - [`bd4887c`](https://github.com/ollama/ollama/commit/bd4887c4ff1a24a46bb5364dc877f0efb9915946) llm: Consistently track unassigned model data - [`d3a7ce2`](https://github.com/ollama/ollama/commit/d3a7ce2e741ffd2f3d7ed394c2348ea8c1222133) llm: Estimate projector memory correctly for Ollama engine ### 📊 Changes **1 file changed** (+28 additions, -20 deletions) <details> <summary>View changed files</summary> 📝 `llm/memory.go` (+28 -20) </details> ### 📄 Description The Llama engine always places vision projectors on the first GPU if one exists. However, the Ollama engine groups it with the output layer, which means the projector is only offloaded if all other layers are offloaded. The memory estimation code always assumes the former layout - this changes it to use the correct layout based on the engine. This addresses two impacts of the current behavior: - In multi-GPU setups, we can crash with OOM errors when we try to allocate memory on a full GPU while another still has space. - If the vision projector is large, it may prevent us from offloading anything when we could have fit some of the text layers. --- <sub>🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.</sub>
GiteaMirror added the pull-request label 2026-04-19 17:15:58 -05:00
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 pull-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#23872
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?