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 #6352] The quality of answer significantly deteriorates after Automatic Quantization #66022

New Issue
Closed
opened 2026-05-03 23:38:31 -05:00 by GiteaMirror · 1 comment
GiteaMirror commented 2026-05-03 23:38:31 -05:00
Owner
Copy Link

Originally created by @garyyang85 on GitHub (Aug 14, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/6352

What is the issue?

Using gemma2-27b for test, after ollama create -q Q8_0, the quality of answer is not very good. The accuracy seems more worse than origin gemma2-9b in hf. What's the principle behind Ollama's quantization? Is it something like post-training quantization? Thanks.

OS

Linux

GPU

Nvidia

CPU

Intel

Ollama version

0.3.5

Originally created by @garyyang85 on GitHub (Aug 14, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/6352 ### What is the issue? Using gemma2-27b for test, after ollama create -q Q8_0, the quality of answer is not very good. The accuracy seems more worse than origin gemma2-9b in hf. What's the principle behind Ollama's quantization? Is it something like post-training quantization? Thanks. ### OS Linux ### GPU Nvidia ### CPU Intel ### Ollama version 0.3.5
GiteaMirror added the bug label 2026-05-03 23:38:31 -05:00
GiteaMirror commented 2026-05-03 23:38:36 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Jan 11, 2026):

It reduces the size of the model by converting fp16/fp32 to smaller data types. This results in a loss of precision and range which affects the quality of the output, much like compressing an image reduces detail and causes artifacts.

<!-- gh-comment-id:3736281168 --> @rick-github commented on GitHub (Jan 11, 2026): It reduces the size of the model by converting fp16/fp32 to smaller data types. This results in a loss of precision and range which affects the quality of the output, much like compressing an image reduces detail and causes artifacts.
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#66022
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?