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

[GH-ISSUE #4825] Adapter doesnt work #28809

New Issue
Closed
opened 2026-04-22 07:21:38 -05:00 by GiteaMirror · 6 comments
GiteaMirror commented 2026-04-22 07:21:38 -05:00
Owner
Copy Link

Originally created by @namogg on GitHub (Jun 5, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/4825

Originally assigned to: @pdevine on GitHub.

What is the issue?

I have converted my adapter to GGML format and create ollama model sucessfully with this comman
ADAPTER downloads/ggml-adapter-model.bin
However when testing the model, the adapter have no affect on the model.

OS

Linux

GPU

Nvidia

CPU

Intel, AMD

Ollama version

0.1.29

Originally created by @namogg on GitHub (Jun 5, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/4825 Originally assigned to: @pdevine on GitHub. ### What is the issue? I have converted my adapter to GGML format and create ollama model sucessfully with this comman ADAPTER downloads/ggml-adapter-model.bin However when testing the model, the adapter have no affect on the model. ### OS Linux ### GPU Nvidia ### CPU Intel, AMD ### Ollama version 0.1.29
GiteaMirror added the bug label 2026-04-22 07:21:39 -05:00
GiteaMirror commented 2026-04-22 07:21:41 -05:00
Author
Owner
Copy Link

@pdevine commented on GitHub (Jun 5, 2024):

@namogg I was just running into the same problem. Can you post the logs just so I can compare with what I've been seeing?
It seems like the adapter gets loaded into memory and then ignored.

<!-- gh-comment-id:2150919317 --> @pdevine commented on GitHub (Jun 5, 2024): @namogg I was just running into the same problem. Can you post the logs just so I can compare with what I've been seeing? It seems like the adapter gets loaded into memory and then ignored.
GiteaMirror commented 2026-04-22 07:21:42 -05:00
Author
Owner
Copy Link

@namogg commented on GitHub (Jun 6, 2024):

nothing interesting with the logs, it just show that the adapter layer is created sucessfully. I haved to work around by merging the adapter with the hugging face model then convert it to GGUF. But i still want to use adapter with ollama.

<!-- gh-comment-id:2151523983 --> @namogg commented on GitHub (Jun 6, 2024): nothing interesting with the logs, it just show that the adapter layer is created sucessfully. I haved to work around by merging the adapter with the hugging face model then convert it to GGUF. But i still want to use adapter with ollama.
GiteaMirror commented 2026-04-22 07:21:42 -05:00
Author
Owner
Copy Link

@pdevine commented on GitHub (Jun 7, 2024):

I've been struggling getting llama.cpp to work w/ any adapters at all. I'd like to be able to allow conversion from both NPZ and Safetensors, which I have the GGLA file being created, but unfortunately I haven't managed to get it to work quite yet. @namogg How are you creating the GGLA file?

<!-- gh-comment-id:2155659735 --> @pdevine commented on GitHub (Jun 7, 2024): I've been struggling getting llama.cpp to work w/ any adapters at all. I'd like to be able to allow conversion from both NPZ and Safetensors, which I have the GGLA file being created, but unfortunately I haven't managed to get it to work quite yet. @namogg How are you creating the GGLA file?
GiteaMirror commented 2026-04-22 07:21:44 -05:00
Author
Owner
Copy Link

@namogg commented on GitHub (Jun 8, 2024):

Nevermind llama.cpp is also too buggy, i used unsloth to train and convert to GGUF sucessfully but it seems the GGUF model doesnt give correct answer. I also tried not using unsloth and using only hugging face but model cannot be merge and load (not enough Vram) therefore cannot be convert to GGUF. I am stuck now

<!-- gh-comment-id:2156081421 --> @namogg commented on GitHub (Jun 8, 2024): Nevermind llama.cpp is also too buggy, i used unsloth to train and convert to GGUF sucessfully but it seems the GGUF model doesnt give correct answer. I also tried not using unsloth and using only hugging face but model cannot be merge and load (not enough Vram) therefore cannot be convert to GGUF. I am stuck now
GiteaMirror commented 2026-04-22 07:21:45 -05:00
Author
Owner
Copy Link

@pdevine commented on GitHub (Jul 9, 2024):

@namogg did you end up getting unstuck at all? I now have the NPZ file conversion working from MLX into GGLAs which can then be loaded. All of this is with LoRAs (no QLoRA yet).

<!-- gh-comment-id:2218669365 --> @pdevine commented on GitHub (Jul 9, 2024): @namogg did you end up getting unstuck at all? I now have the NPZ file conversion working from MLX into GGLAs which can then be loaded. All of this is with LoRAs (no QLoRA yet).
GiteaMirror commented 2026-04-22 07:21:45 -05:00
Author
Owner
Copy Link

@pdevine commented on GitHub (Jul 10, 2024):

Let's track this in #4618 .

EDIT: On second thought, let's track this in #156

<!-- gh-comment-id:2221177724 --> @pdevine commented on GitHub (Jul 10, 2024): Let's track this in #4618 . EDIT: On second thought, let's track this in #156
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#28809
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?