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 #10749] Regression in 0.7.0 : can't quantize to Q2_K #7061

New Issue
Closed
opened 2026-04-12 18:58:43 -05:00 by GiteaMirror · 3 comments
GiteaMirror commented 2026-04-12 18:58:43 -05:00
Owner
Copy Link

Originally created by @lowlyocean on GitHub (May 16, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/10749

What is the issue?

This was working in 0.6.x but now seems to be unsupported in latest version

Relevant log output

Error: unsupported quantization type Q2_K - supported types are F32, F16, Q4_K_S, Q4_K_M, Q8_0

OS

No response

GPU

No response

CPU

No response

Ollama version

0.7.0

Originally created by @lowlyocean on GitHub (May 16, 2025). Original GitHub issue: https://github.com/ollama/ollama/issues/10749 ### What is the issue? This was working in 0.6.x but now seems to be unsupported in latest version ### Relevant log output ```shell Error: unsupported quantization type Q2_K - supported types are F32, F16, Q4_K_S, Q4_K_M, Q8_0 ``` ### OS _No response_ ### GPU _No response_ ### CPU _No response_ ### Ollama version 0.7.0
GiteaMirror added the bug label 2026-04-12 18:58:43 -05:00
GiteaMirror commented 2026-04-12 18:58:44 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (May 16, 2025):

ollama only supports Q4 and Q8 quantizations during model creation.

<!-- gh-comment-id:2887606781 --> @rick-github commented on GitHub (May 16, 2025): ollama [only supports](https://github.com/ollama/ollama/pull/10647#issuecomment-2873563847) Q4 and Q8 quantizations during model creation.
GiteaMirror commented 2026-04-12 18:58:44 -05:00
Author
Owner
Copy Link

@lowlyocean commented on GitHub (May 16, 2025):

That's not sufficiently small for my use case. per the person's response to your question on that thread, it sounds like we will be phased out of using Q2_k over time (even though initially, we can get away with using llama.cpp for quantization). That's disappointing

<!-- gh-comment-id:2887610982 --> @lowlyocean commented on GitHub (May 16, 2025): That's not sufficiently small for my use case. per the person's response to your question on that thread, it sounds like we will be phased out of using Q2_k over time (even though initially, we can get away with using llama.cpp for quantization). That's disappointing
GiteaMirror commented 2026-04-12 18:58:45 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (May 16, 2025):

There are other projects that will likely continue to support quantization levels that ollama doesn't. You probably know these already: llama.cpp (of course), LMStudio, Mistral.rs. vLLM support GGUF and while I don't know if they support very low quant levels, they have their own LLM compressor.

<!-- gh-comment-id:2887645305 --> @rick-github commented on GitHub (May 16, 2025): There are other projects that will likely continue to support quantization levels that ollama doesn't. You probably know these already: [llama.cpp](https://github.com/ggml-org/llama.cpp) (of course), [LMStudio](https://lmstudio.ai/), [Mistral.rs](https://github.com/EricLBuehler/mistral.rs). [vLLM](https://github.com/vllm-project/vllm) support GGUF and while I don't know if they support very low quant levels, they have their own [LLM compressor](https://github.com/vllm-project/llm-compressor).
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#7061
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?