github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-08 00:51:34 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[PR #9557] [MERGED] ollamarunner: Improve multimodal input handling #75293

New Issue
Closed
opened 2026-05-05 07:43:56 -05:00 by GiteaMirror · 0 comments
GiteaMirror commented 2026-05-05 07:43:56 -05:00
Owner
Copy Link

📋 Pull Request Information

Original PR: https://github.com/ollama/ollama/pull/9557
Author: @jessegross
Created: 3/6/2025
Status: ✅ Merged
Merged: 3/7/2025
Merged by: @jessegross

Base: main ← Head: jessegross/mulitmodal

📝 Commits (2)

  • d771a00 model: Don't unconditionally add special tokens
  • 2948c0a ollamarunner: Improve multimodal input handling

📊 Changes

8 files changed (+254 additions, -137 deletions)

View changed files

📝 llm/server.go (+1 -1)
📝 model/model.go (+63 -7)
📝 model/models/mllama/model.go (+72 -29)
📝 model/process_text.go (+3 -3)
📝 model/process_text_test.go (+7 -7)
📝 runner/ollamarunner/cache.go (+8 -12)
📝 runner/ollamarunner/cache_test.go (+41 -33)
📝 runner/ollamarunner/runner.go (+59 -45)

📄 Description

Various vision models have different requirements for how they receive their inputs. For example:

  • Mllama wants images together with text and the image embeddings don't themselves have positions or get stored in the main KV cache
  • Llava-style models feed in embeddings similar to tokens and images correspond to a varying number of tokens in the cache.

In addition, the strategy for providing inputs must support batching and multiple sequences, which are managed by the runner. At the same time, we want to keep data handling fully in the model so that new architectures are not bottlenecked by runner code which does not understand their particular requirements.

This provides a method for models to edit the input stream so that it meets their needs while still being in a format that the runner understands. This allows the runner to avoid special processing for different models.

In addition, this fixes a regression where non-vision models may try to incorrectly interpret images.

🔄 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/9557 **Author:** [@jessegross](https://github.com/jessegross) **Created:** 3/6/2025 **Status:** ✅ Merged **Merged:** 3/7/2025 **Merged by:** [@jessegross](https://github.com/jessegross) **Base:** `main` ← **Head:** `jessegross/mulitmodal` --- ### 📝 Commits (2) - [`d771a00`](https://github.com/ollama/ollama/commit/d771a00438adee83f1fef1d91101c62e03e77928) model: Don't unconditionally add special tokens - [`2948c0a`](https://github.com/ollama/ollama/commit/2948c0a881ae525bf07f32e9407aeb2c659b89ee) ollamarunner: Improve multimodal input handling ### 📊 Changes **8 files changed** (+254 additions, -137 deletions) <details> <summary>View changed files</summary> 📝 `llm/server.go` (+1 -1) 📝 `model/model.go` (+63 -7) 📝 `model/models/mllama/model.go` (+72 -29) 📝 `model/process_text.go` (+3 -3) 📝 `model/process_text_test.go` (+7 -7) 📝 `runner/ollamarunner/cache.go` (+8 -12) 📝 `runner/ollamarunner/cache_test.go` (+41 -33) 📝 `runner/ollamarunner/runner.go` (+59 -45) </details> ### 📄 Description Various vision models have different requirements for how they receive their inputs. For example: - Mllama wants images together with text and the image embeddings don't themselves have positions or get stored in the main KV cache - Llava-style models feed in embeddings similar to tokens and images correspond to a varying number of tokens in the cache. In addition, the strategy for providing inputs must support batching and multiple sequences, which are managed by the runner. At the same time, we want to keep data handling fully in the model so that new architectures are not bottlenecked by runner code which does not understand their particular requirements. This provides a method for models to edit the input stream so that it meets their needs while still being in a format that the runner understands. This allows the runner to avoid special processing for different models. In addition, this fixes a regression where non-vision models may try to incorrectly interpret images. --- <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 2026-05-05 07:43:56 -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#75293
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?