github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-07 00:22:43 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #9190] When I run any model using 'ollama run', there is a chance it will output a series of '@' characters. #52500

New Issue
Closed
opened 2026-04-28 23:29:47 -05:00 by GiteaMirror · 5 comments
GiteaMirror commented 2026-04-28 23:29:47 -05:00
Owner
Copy Link

Originally created by @HanShengGoodWay on GitHub (Feb 18, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/9190

What is the issue?

I'm not sure what happened.
When I use ollama run with any model, it sometimes outputs "@" after a while, regardless of the input.
To be precise, it prints "@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@", which is exactly 31 "@" characters.

I'm using
Windows 11
RTX 4090
CUDA version: 12.2

How should I start debugging this issue?

Relevant log output

ollama run  deepseek-r1:14B
>>> hi
<think>

</think>

Hello! How can I assist you today? 😊

>>> hi
<think>
Alright, the user just sent "hi" again. They might be trying to get my attention or test if I'm responsive.

I should respond in a friendly and welcoming manner to make them feel comfortable.

Maybe I'll greet them back and offer help with a smiley face to keep it warm.

Keeping it open-ended so they know I'm here to assist with whatever they need.
</think>

Hi! 😊 How can I assist you today?

>>> hi
<think>
Alright, the user is saying "hi@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@

>>> hi
@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@

>>> hi
@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@

OS

Windows

GPU

Nvidia

CPU

Intel

Ollama version

0.5.11

Originally created by @HanShengGoodWay on GitHub (Feb 18, 2025). Original GitHub issue: https://github.com/ollama/ollama/issues/9190 ### What is the issue? I'm not sure what happened. When I use ollama run with any model, it sometimes outputs "@" after a while, regardless of the input. To be precise, it prints "@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@", which is exactly 31 "@" characters. I'm using Windows 11 RTX 4090 CUDA version: 12.2 How should I start debugging this issue? ### Relevant log output ```shell ollama run deepseek-r1:14B >>> hi <think> </think> Hello! How can I assist you today? 😊 >>> hi <think> Alright, the user just sent "hi" again. They might be trying to get my attention or test if I'm responsive. I should respond in a friendly and welcoming manner to make them feel comfortable. Maybe I'll greet them back and offer help with a smiley face to keep it warm. Keeping it open-ended so they know I'm here to assist with whatever they need. </think> Hi! 😊 How can I assist you today? >>> hi <think> Alright, the user is saying "hi@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@ >>> hi @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@ >>> hi @@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@ ``` ### OS Windows ### GPU Nvidia ### CPU Intel ### Ollama version 0.5.11
GiteaMirror added the bug label 2026-04-28 23:29:47 -05:00
GiteaMirror commented 2026-04-28 23:30:49 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Feb 18, 2025):

Server logs may aid in debugging.

<!-- gh-comment-id:2664995094 --> @rick-github commented on GitHub (Feb 18, 2025): [Server logs](https://github.com/ollama/ollama/blob/main/docs/troubleshooting.md#how-to-troubleshoot-issues) may aid in debugging.
GiteaMirror commented 2026-04-28 23:30:50 -05:00
Author
Owner
Copy Link

@HanShengGoodWay commented on GitHub (Feb 19, 2025):

Server logs may aid in debugging.

I observed the output using ollama serve (should this be consistent with the logs?).

I have seen this error before (I asked GPT to format it for better readability), but it does not always appear whenever a series of "@" symbols is present.

[2025-02-19T15:37:55.782+08:00] [WARN] [gpu.go:434]
   unable to get device handle GPU-e6f1cdee-e5db-49ee-1fcf-55aa681b3665: 999
   msg: "error looking up nvidia GPU memory"

[2025-02-19T15:38:00.782+08:00] [WARN] [sched.go:646]
   msg: "gpu VRAM usage didn't recover within timeout"
   seconds: 5.0016686
   model: C:\Users\HS\.ollama\models\blobs\sha256-e2f46f5b501c2982b2c495a4694cb4e620aabfa2c37ebb23a90ffc8cce93854b

Typically, I only see the usual model loading information and the requests I send. However, at some point, during one of my requests, a series of "@" suddenly starts appearing, and without any other error messages.

From my observations, this issue seems more likely to occur when switching between two models. However, I cannot guarantee that it only happens when unloading Model A and loading Model B.

I really appreciate your help and suggestions!

<!-- gh-comment-id:2667924452 --> @HanShengGoodWay commented on GitHub (Feb 19, 2025): > [Server logs](https://github.com/ollama/ollama/blob/main/docs/troubleshooting.md#how-to-troubleshoot-issues) may aid in debugging. I observed the output using ollama serve (should this be consistent with the logs?). I have seen this error before (I asked GPT to format it for better readability), but it does not always appear whenever a series of "@" symbols is present. ``` [2025-02-19T15:37:55.782+08:00] [WARN] [gpu.go:434] unable to get device handle GPU-e6f1cdee-e5db-49ee-1fcf-55aa681b3665: 999 msg: "error looking up nvidia GPU memory" [2025-02-19T15:38:00.782+08:00] [WARN] [sched.go:646] msg: "gpu VRAM usage didn't recover within timeout" seconds: 5.0016686 model: C:\Users\HS\.ollama\models\blobs\sha256-e2f46f5b501c2982b2c495a4694cb4e620aabfa2c37ebb23a90ffc8cce93854b ``` Typically, I only see the usual model loading information and the requests I send. However, at some point, during one of my requests, a series of "@" suddenly starts appearing, and without any other error messages. From my observations, this issue seems more likely to occur when switching between two models. However, I cannot guarantee that it only happens when unloading Model A and loading Model B. I really appreciate your help and suggestions!
GiteaMirror commented 2026-04-28 23:30:52 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Feb 19, 2025):

#8235 had a similar problem with repeating '@'s, running a GPU VRAM tester revealed a faulty card.

<!-- gh-comment-id:2668019889 --> @rick-github commented on GitHub (Feb 19, 2025): #8235 had a similar problem with repeating '@'s, running a [GPU VRAM tester]( https://www.programming4beginners.com/gpumemtest) revealed a faulty card.
GiteaMirror commented 2026-04-28 23:30:54 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Feb 27, 2025):

Another user with this problem (#8495) found that upgrading the Nvidia driver to 572.60 resolved their problem.

<!-- gh-comment-id:2689059462 --> @rick-github commented on GitHub (Feb 27, 2025): Another user with this problem (#8495) found that upgrading the Nvidia driver to [572.60](https://www.nvidia.com/en-us/geforce/drivers/results/241091/) resolved their problem.
GiteaMirror commented 2026-04-28 23:31:01 -05:00
Author
Owner
Copy Link

@HanShengGoodWay commented on GitHub (Apr 17, 2025):

另一個遇到此問題的用戶（#8495）發現將 Nvidia 驅動程式升級到572.60可以解決他們的問題。

Later I found that the problem was caused by ollama not being updated. When these abnormal outputs were found, updating ollama (maybe restarting the computer?) could solve the problem.

<!-- gh-comment-id:2811518954 --> @HanShengGoodWay commented on GitHub (Apr 17, 2025): > 另一個遇到此問題的用戶（[#8495](https://github.com/ollama/ollama/issues/8495)）發現將 Nvidia 驅動程式升級到[572.60](https://www.nvidia.com/en-us/geforce/drivers/results/241091/)可以解決他們的問題。 Later I found that the problem was caused by ollama not being updated. When these abnormal outputs were found, updating ollama (maybe restarting the computer?) could solve the problem.
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#52500
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?