github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-07 16:40:08 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #6385] Significant Drop in Prompt Adherence in Updated Gemma2 Model #4011

New Issue
Closed
opened 2026-04-12 14:52:56 -05:00 by GiteaMirror · 1 comment
GiteaMirror commented 2026-04-12 14:52:56 -05:00
Owner
Copy Link

Originally created by @shzhou12 on GitHub (Aug 16, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/6385

What is the issue?

I recently noticed that the Gemma2 model was updated 5 weeks ago, resulting in a new version of gemma2:9b-instruct-fp16:

  • Older Version (6 weeks ago): gemma2:9b-instruct-fp16 - 9de55d4bf6ae - 18 GB
  • Updated Version (5 weeks ago): gemma2:9b-instruct-fp16 - 28e6684b0850 - 18 GB

After switching to the updated version 28e6684b0850, I've observed a significant decrease in the model's ability to adhere to prompts in my specific downstream tasks.

Could you please clarify why there are two different versions of gemma2:9b-instruct-fp16 and what changes were made between these versions? Should I revert to the older version 9de55d4bf6ae to maintain the previous performance?

OS

macOS

GPU

Apple

CPU

Apple

Ollama version

0.2.8

Originally created by @shzhou12 on GitHub (Aug 16, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/6385 ### What is the issue? I recently noticed that the Gemma2 model was updated 5 weeks ago, resulting in a new version of gemma2:9b-instruct-fp16: - Older Version (6 weeks ago): gemma2:9b-instruct-fp16 - **9de55d4bf6ae** - 18 GB - Updated Version (5 weeks ago): gemma2:9b-instruct-fp16 - **28e6684b0850** - 18 GB After switching to the updated version 28e6684b0850, I've observed a significant decrease in the model's ability to adhere to prompts in my specific downstream tasks. Could you please clarify why there are two different versions of gemma2:9b-instruct-fp16 and what changes were made between these versions? Should I revert to the older version 9de55d4bf6ae to maintain the previous performance? ### OS macOS ### GPU Apple ### CPU Apple ### Ollama version 0.2.8
GiteaMirror added the bug label 2026-04-12 14:52:56 -05:00
GiteaMirror commented 2026-04-12 14:52:58 -05:00
Author
Owner
Copy Link

@igorschlum commented on GitHub (Aug 16, 2024):

Hi @shzhou12 thank you for your input. Do you have an exemple of prompt where performance dropped?

Why don't you update to the latest version of ollama and stay with 0.2.8?

<!-- gh-comment-id:2292857941 --> @igorschlum commented on GitHub (Aug 16, 2024): Hi @shzhou12 thank you for your input. Do you have an exemple of prompt where performance dropped? Why don't you update to the latest version of ollama and stay with 0.2.8?
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#4011
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?