github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-06 08:02:14 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #7289] VPTQ Model Quantization Support in Ollama #51144

New Issue
Open
opened 2026-04-28 18:33:15 -05:00 by GiteaMirror · 4 comments
GiteaMirror commented 2026-04-28 18:33:15 -05:00
Owner
Copy Link

Originally created by @YangWang92 on GitHub (Oct 21, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/7289

Hi all,

We recently developed a fully open-source quantization method called VPTQ (Vector Post-Training Quantization) https://github.com/microsoft/VPTQ which enables fast quantization of large language models (LLMs) down to 1-4 bits. The community has also helped release several models using this method https://huggingface.co/VPTQ-community. I am personally very interested in integrating VPTQ into ollama/llama.cpp.

One of the key advantages of VPTQ is that the dequantization method is very straightforward, relying only on a simple lookup table.

I would like to ask for guidance on how best to support this quantization method within Ollama, even if it's on my own fork. Specifically, which approach should I take?

  1. Define a series of new models (e.g., vptq-llama3.1) using existing data types (int32, fp16), and hide the model dequantization within a separate dequant op.

  2. Define a new quantization data type (e.g., a custom lookup table data structure)?

I’d love to hear your thoughts or any suggestions on how to proceed!

Thank you!
Yang

Originally created by @YangWang92 on GitHub (Oct 21, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/7289 Hi all, We recently developed a fully open-source quantization method called VPTQ (Vector Post-Training Quantization) [https://github.com/microsoft/VPTQ](https://github.com/microsoft/VPTQ) which enables fast quantization of large language models (LLMs) down to 1-4 bits. The community has also helped release several models using this method [https://huggingface.co/VPTQ-community](https://huggingface.co/VPTQ-community). I am personally very interested in integrating VPTQ into ollama/llama.cpp. One of the key advantages of VPTQ is that the dequantization method is very straightforward, relying only on a simple lookup table. I would like to ask for guidance on how best to support this quantization method within Ollama, even if it's on my own fork. Specifically, which approach should I take? 1. Define a series of new models (e.g., vptq-llama3.1) using existing data types (int32, fp16), and hide the model dequantization within a separate dequant op. 2. Define a new quantization data type (e.g., a custom lookup table data structure)? I’d love to hear your thoughts or any suggestions on how to proceed! Thank you! Yang
GiteaMirror added the feature request label 2026-04-28 18:33:15 -05:00
GiteaMirror commented 2026-04-28 18:34:09 -05:00
Author
Owner
Copy Link

@guibirow commented on GitHub (Oct 25, 2024):

#2821

<!-- gh-comment-id:2437418374 --> @guibirow commented on GitHub (Oct 25, 2024): #2821
GiteaMirror commented 2026-04-28 18:34:10 -05:00
Author
Owner
Copy Link

@YangWang92 commented on GitHub (Oct 25, 2024):

#2821
Hi @guibirow,
cool, thanks for point this out!

<!-- gh-comment-id:2437421986 --> @YangWang92 commented on GitHub (Oct 25, 2024): > #2821 Hi @guibirow, cool, thanks for point this out!
GiteaMirror commented 2026-04-28 18:34:11 -05:00
Author
Owner
Copy Link

@guibirow commented on GitHub (Oct 25, 2024):

How does this compare/relate to https://github.com/microsoft/BitNet

<!-- gh-comment-id:2437430858 --> @guibirow commented on GitHub (Oct 25, 2024): How does this compare/relate to https://github.com/microsoft/BitNet
GiteaMirror commented 2026-04-28 18:34:12 -05:00
Author
Owner
Copy Link

@YangWang92 commented on GitHub (Oct 25, 2024):

How does this compare/relate to https://github.com/microsoft/BitNet

Our VPTQ is a weight-only quantization method, so it clearly cannot achieve the same results as Bitnet at the same bit level. The advantage of VPTQ is that it provides 1-2 bit quantization for the latest models, such as Llama 3.1 and Qwen2.5. https://huggingface.co/VPTQ-community

<!-- gh-comment-id:2437441386 --> @YangWang92 commented on GitHub (Oct 25, 2024): > How does this compare/relate to https://github.com/microsoft/BitNet Our VPTQ is a weight-only quantization method, so it clearly cannot achieve the same results as Bitnet at the same bit level. The advantage of VPTQ is that it provides 1-2 bit quantization for the latest models, such as Llama 3.1 and Qwen2.5. https://huggingface.co/VPTQ-community
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 feature 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#51144
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?