github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-07-17 13:47:08 -05:00
Code Issues 16.9k Packages Projects Releases 122 Wiki Activity

[GH-ISSUE #15573] Infinity loop in tool calls with Gemma4 + Ollama 0.20.6+ #72001

New Issue
Open
opened 2026-05-05 03:16:50 -05:00 by GiteaMirror · 3 comments
GiteaMirror commented 2026-05-05 03:16:50 -05:00
Owner
Copy Link

Originally created by @ByCzech on GitHub (Apr 14, 2026).
Original GitHub issue: https://github.com/ollama/ollama/issues/15573

What is the issue?

Upgrade Ollama to version 0.20.6 (edit and 0.20.7 too) causing infinity loop in tool calls with gemma4:31b. After reverting back to version 0.20.5 this problem will disappear. Observed only with the gemma4 model, other models e.g. qwen3.5 work with tool calls without problems.

Relevant log output

OS

Linux

GPU

AMD

CPU

AMD

Ollama version

0.20.6 / 0.20.7 / 0.21.0

Originally created by @ByCzech on GitHub (Apr 14, 2026). Original GitHub issue: https://github.com/ollama/ollama/issues/15573 ### What is the issue? Upgrade Ollama to version 0.20.6 (edit and 0.20.7 too) causing infinity loop in tool calls with gemma4:31b. After reverting back to version 0.20.5 this problem will disappear. Observed only with the gemma4 model, other models e.g. qwen3.5 work with tool calls without problems. ### Relevant log output ```shell ``` ### OS Linux ### GPU AMD ### CPU AMD ### Ollama version 0.20.6 / 0.20.7 / 0.21.0
GiteaMirror added the bug label 2026-05-05 03:16:50 -05:00
GiteaMirror commented 2026-05-05 03:16:54 -05:00
Author
Owner
Copy Link

@PureBlissAK commented on GitHub (Apr 18, 2026):

🤖 Automated Triage & Analysis Report

Issue: #15573
Analyzed: 2026-04-18T18:20:40.963033

Analysis

  • Type: unknown
  • Severity: medium
  • Components: unknown

Implementation Plan

  • Effort: medium
  • Steps:

This issue has been triaged and marked for implementation.

<!-- gh-comment-id:4274306969 --> @PureBlissAK commented on GitHub (Apr 18, 2026): <!-- ollama-issue-orchestrator:v1 issue:15573 --> ## 🤖 Automated Triage & Analysis Report **Issue**: #15573 **Analyzed**: 2026-04-18T18:20:40.963033 ### Analysis - **Type**: unknown - **Severity**: medium - **Components**: unknown ### Implementation Plan - **Effort**: medium - **Steps**: *This issue has been triaged and marked for implementation.*
GiteaMirror commented 2026-05-05 03:16:55 -05:00
Author
Owner
Copy Link

@noochboy commented on GitHub (Apr 28, 2026):

Encountering the same issues here.

both with qwen3.5 and gemma4

cloud models are ok

Just tried with qwen3.6:27b as well...

● let's try again... What were we researching? Please check the TASKS.md file and the FIELDS.md file for context 
 Let me read those files to get back up to speed.

 <command_bash>
 cat TASKS.md</command_bash>

another test

● can you list the content of file /home/noochboy/.hermes/profiles/researcher/FIELDS.md 
 <antThinking>
 The user wants to see the content of a file. I should read it.
 </antThinking>

 <antThinking>
 The user wants to view the content of a specific file. I'll use cat to display it.
 </antThinking>

 bash
 cat /home/noochboy/.hermes/profiles/researcher/FIELDS.md
<!-- gh-comment-id:4333725238 --> @noochboy commented on GitHub (Apr 28, 2026): Encountering the same issues here. both with qwen3.5 and gemma4 cloud models are ok Just tried with qwen3.6:27b as well... <prompt> ● let's try again... What were we researching? Please check the TASKS.md file and the FIELDS.md file for context </prompt> <reply> Let me read those files to get back up to speed. <command_bash> cat TASKS.md</command_bash> </reply> another test <prompt> ● can you list the content of file /home/noochboy/.hermes/profiles/researcher/FIELDS.md </prompt> <reply> <antThinking> The user wants to see the content of a file. I should read it. </antThinking> <antThinking> The user wants to view the content of a specific file. I'll use cat to display it. </antThinking> bash cat /home/noochboy/.hermes/profiles/researcher/FIELDS.md </reply>
GiteaMirror commented 2026-05-05 03:16:58 -05:00
Author
Owner
Copy Link

@ByCzech commented on GitHub (May 4, 2026):

Can't confirm that behavior in other model than Gemma4. Other models i.e. Qwen3.x, Mistral etc. works well for me on Vulkan iGPU, GPU and CUDA GPU.

<!-- gh-comment-id:4373882155 --> @ByCzech commented on GitHub (May 4, 2026): Can't confirm that behavior in other model than Gemma4. Other models i.e. Qwen3.x, Mistral etc. works well for me on Vulkan iGPU, GPU and CUDA GPU.
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#72001
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?