github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-06 16:11:34 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[PR #290] [MERGED] implement loading ggml lora adapters through the modelfile #35954

New Issue
Closed
opened 2026-04-22 20:41:40 -05:00 by GiteaMirror · 0 comments
GiteaMirror commented 2026-04-22 20:41:40 -05:00
Owner
Copy Link

📋 Pull Request Information

Original PR: https://github.com/ollama/ollama/pull/290
Author: @mxyng
Created: 8/5/2023
Status: ✅ Merged
Merged: 8/11/2023
Merged by: @mxyng

Base: main ← Head: add-adapter-layers

📝 Commits (2)

  • 6de5d03 implement loading ggml lora adapters through the modelfile
  • 37c9a8e add lora docs

📊 Changes

6 files changed (+75 additions, -14 deletions)

View changed files

📝 docs/modelfile.md (+10 -1)
📝 llm/llama.go (+16 -1)
📝 llm/llm.go (+2 -2)
📝 parser/parser.go (+1 -1)
📝 server/images.go (+45 -8)
📝 server/routes.go (+1 -1)

📄 Description

LoRA adapters can be added to Ollama models through the Modelfile and automatically applied when the model is loaded:

FROM llama2:13b
TEMPLATE {{ .Prompt }}
ADAPTER ./llama2-13b-storywriter-lora.ggml.bin

A few caveats:

  • LoRA adapters must be GGML. If the adapter isn't GGML, it can be converted with the convert-lora-to-ggml.py script in https://github.com/ggerganov/llama.cpp
  • Using adapters with quantized weights might not produce good results
  • Using adapters disables mmap
  • It's possible to apply multiple adapters but
    1. Ordering is important
    2. There may be unintended side effects since most adapters are not intended to be layered on other adapters
    3. Performance may degrade with more adapters

🔄 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/290 **Author:** [@mxyng](https://github.com/mxyng) **Created:** 8/5/2023 **Status:** ✅ Merged **Merged:** 8/11/2023 **Merged by:** [@mxyng](https://github.com/mxyng) **Base:** `main` ← **Head:** `add-adapter-layers` --- ### 📝 Commits (2) - [`6de5d03`](https://github.com/ollama/ollama/commit/6de5d032e1da8f7a2a15010a5422cbddc5d0b932) implement loading ggml lora adapters through the modelfile - [`37c9a8e`](https://github.com/ollama/ollama/commit/37c9a8eea98626bb32b05ec2a82279082317f0d6) add lora docs ### 📊 Changes **6 files changed** (+75 additions, -14 deletions) <details> <summary>View changed files</summary> 📝 `docs/modelfile.md` (+10 -1) 📝 `llm/llama.go` (+16 -1) 📝 `llm/llm.go` (+2 -2) 📝 `parser/parser.go` (+1 -1) 📝 `server/images.go` (+45 -8) 📝 `server/routes.go` (+1 -1) </details> ### 📄 Description LoRA adapters can be added to Ollama models through the Modelfile and automatically applied when the model is loaded: ``` FROM llama2:13b TEMPLATE {{ .Prompt }} ADAPTER ./llama2-13b-storywriter-lora.ggml.bin ``` A few caveats: * LoRA adapters must be GGML. If the adapter isn't GGML, it can be converted with the `convert-lora-to-ggml.py` script in https://github.com/ggerganov/llama.cpp * Using adapters with quantized weights might not produce good results * Using adapters disables mmap * It's possible to apply multiple adapters but 1. Ordering is important 2. There may be unintended side effects since most adapters are not intended to be layered on other adapters 3. Performance may degrade with more adapters --- <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 2026-04-22 20:41:40 -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#35954
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?