github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-07-11 07:32:07 -05:00
Code Issues 16.9k Packages Projects Releases 122 Wiki Activity

[GH-ISSUE #4890] qwen2 not run correctly #28846

New Issue
Closed
opened 2026-04-22 07:24:02 -05:00 by GiteaMirror · 23 comments
GiteaMirror commented 2026-04-22 07:24:02 -05:00
Owner
Copy Link

Originally created by @FreemanFeng on GitHub (Jun 7, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/4890

What is the issue?

ollama run qwen2

tell a story
GG

OnceGGG thereGGGGGGGGGGGG wasGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG

讲一个故事
GG

GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG

Above is what qwen2 model response, not sure if it is ollama not supporting qwen2 yet or not?

OS

Windows

GPU

Nvidia

CPU

Intel

Ollama version

0.1.41

Originally created by @FreemanFeng on GitHub (Jun 7, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/4890 ### What is the issue? ollama run qwen2 >>> tell a story GG OnceGGG thereGGGGGGGGGGGG wasGGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG >>> 讲一个故事 GG GGGGGGGGGGGGGGGGGGGGGGGGGGGGGGG >>> Above is what qwen2 model response, not sure if it is ollama not supporting qwen2 yet or not? ### OS Windows ### GPU Nvidia ### CPU Intel ### Ollama version 0.1.41
GiteaMirror added the bug label 2026-04-22 07:24:02 -05:00
GiteaMirror commented 2026-04-22 07:24:03 -05:00
Author
Owner
Copy Link

@slash130 commented on GitHub (Jun 7, 2024):

在linux服务器上遇到相同的问题，在macmini上是正常的。

<!-- gh-comment-id:2153735697 --> @slash130 commented on GitHub (Jun 7, 2024): 在linux服务器上遇到相同的问题，在macmini上是正常的。
GiteaMirror commented 2026-04-22 07:24:04 -05:00
Author
Owner
Copy Link

@tonightmare commented on GitHub (Jun 7, 2024):

I meet the same problem with 4090 in ubuntu linux platform

<!-- gh-comment-id:2153735792 --> @tonightmare commented on GitHub (Jun 7, 2024): I meet the same problem with 4090 in ubuntu linux platform
GiteaMirror commented 2026-04-22 07:24:04 -05:00
Author
Owner
Copy Link

@freeng commented on GitHub (Jun 7, 2024):

I meet the same problem
3060 6g
16g ram
windows 11

<!-- gh-comment-id:2153740671 --> @freeng commented on GitHub (Jun 7, 2024): I meet the same problem 3060 6g 16g ram windows 11
GiteaMirror commented 2026-04-22 07:24:05 -05:00
Author
Owner
Copy Link

@Moeblack commented on GitHub (Jun 7, 2024):

I meet the same problem on wsl2. with 4090 and 5800x3d 64G ram
9b330dc35f2c203d8278807de423842c

<!-- gh-comment-id:2153741898 --> @Moeblack commented on GitHub (Jun 7, 2024): I meet the same problem on wsl2. with 4090 and 5800x3d 64G ram ![9b330dc35f2c203d8278807de423842c](https://github.com/ollama/ollama/assets/28835093/4912f714-8223-4131-ab3b-743d540846b0)
GiteaMirror commented 2026-04-22 07:24:05 -05:00
Author
Owner
Copy Link

@ajie425 commented on GitHub (Jun 7, 2024):

win11上也出现这种问题
image

<!-- gh-comment-id:2153749465 --> @ajie425 commented on GitHub (Jun 7, 2024): win11上也出现这种问题 ![image](https://github.com/ollama/ollama/assets/40158391/a8102510-fa5c-4d9b-b1a1-b64c228f9e4f)
GiteaMirror commented 2026-04-22 07:24:06 -05:00
Author
Owner
Copy Link

@jingyibo123 commented on GitHub (Jun 7, 2024):

Same on linux,

Also, similar bug with qwen_v1.5 when num_ctx >2048, if it's related

<!-- gh-comment-id:2153761783 --> @jingyibo123 commented on GitHub (Jun 7, 2024): Same on linux, Also, similar bug with `qwen_v1.5` when `num_ctx >2048`, if it's related
GiteaMirror commented 2026-04-22 07:24:06 -05:00
Author
Owner
Copy Link

@Leroy-X commented on GitHub (Jun 7, 2024):

GGGGG +1

<!-- gh-comment-id:2153775999 --> @Leroy-X commented on GitHub (Jun 7, 2024): GGGGG +1
GiteaMirror commented 2026-04-22 07:24:07 -05:00
Author
Owner
Copy Link

@Forevery1 commented on GitHub (Jun 7, 2024):

一样输出一堆gggggggg

<!-- gh-comment-id:2153808053 --> @Forevery1 commented on GitHub (Jun 7, 2024): 一样输出一堆gggggggg
GiteaMirror commented 2026-04-22 07:24:07 -05:00
Author
Owner
Copy Link

@Tining123 commented on GitHub (Jun 7, 2024):

./ollama --version
ollama version is 0.1.41

same issue

<!-- gh-comment-id:2153836463 --> @Tining123 commented on GitHub (Jun 7, 2024): ./ollama --version ollama version is 0.1.41 same issue
GiteaMirror commented 2026-04-22 07:24:07 -05:00
Author
Owner
Copy Link

@lingyezhixing commented on GitHub (Jun 7, 2024):

same question

<!-- gh-comment-id:2153856116 --> @lingyezhixing commented on GitHub (Jun 7, 2024): same question
GiteaMirror commented 2026-04-22 07:24:08 -05:00
Author
Owner
Copy Link

@HorseLuke commented on GitHub (Jun 7, 2024):

Seems that if it is run by CPU only, result is normal but run slowly. Using GPU will result such problem.

<!-- gh-comment-id:2153910820 --> @HorseLuke commented on GitHub (Jun 7, 2024): Seems that if it is run by CPU only, result is normal but run slowly. Using GPU will result such problem.
GiteaMirror commented 2026-04-22 07:24:08 -05:00
Author
Owner
Copy Link

@catnaut commented on GitHub (Jun 7, 2024):

same question

<!-- gh-comment-id:2154131718 --> @catnaut commented on GitHub (Jun 7, 2024): same question
GiteaMirror commented 2026-04-22 07:24:09 -05:00
Author
Owner
Copy Link

@MaxJa4 commented on GitHub (Jun 7, 2024):

Had the same issue. After fully closing/stopping Ollama (not just /bye in the console) and restarting it, everything worked again.

<!-- gh-comment-id:2154158227 --> @MaxJa4 commented on GitHub (Jun 7, 2024): Had the same issue. After fully closing/stopping Ollama (not just /bye in the console) and restarting it, everything worked again.
GiteaMirror commented 2026-04-22 07:24:09 -05:00
Author
Owner
Copy Link

@ghost commented on GitHub (Jun 7, 2024):

i just added this parameter OLLAMA_FLASH_ATTENTION=1 when running ollama serve, and qwen2 worked on my gpu

<!-- gh-comment-id:2154223293 --> @ghost commented on GitHub (Jun 7, 2024): i just added this parameter OLLAMA_FLASH_ATTENTION=1 when running ollama serve, and qwen2 worked on my gpu
GiteaMirror commented 2026-04-22 07:24:09 -05:00
Author
Owner
Copy Link

@Moeblack commented on GitHub (Jun 7, 2024):

i just added this parameter OLLAMA_FLASH_ATTENTION=1 when running ollama serve, and qwen2 worked on my gpu我刚刚在运行 ollama 服务时添加了 OLLAMA_FLASH_ATTENTION=1 参数，结果 qwen2 在我的显卡上正常运行了。

well, succeed.
image

<!-- gh-comment-id:2154246924 --> @Moeblack commented on GitHub (Jun 7, 2024): > i just added this parameter OLLAMA_FLASH_ATTENTION=1 when running ollama serve, and qwen2 worked on my gpu我刚刚在运行 ollama 服务时添加了 OLLAMA_FLASH_ATTENTION=1 参数，结果 qwen2 在我的显卡上正常运行了。 well, succeed. ![image](https://github.com/ollama/ollama/assets/28835093/b97c27f2-ecbb-4cea-97f8-c7931fa584d6)
GiteaMirror commented 2026-04-22 07:24:10 -05:00
Author
Owner
Copy Link

@jiangweiatgithub commented on GitHub (Jun 7, 2024):

This works for me!

<!-- gh-comment-id:2154306781 --> @jiangweiatgithub commented on GitHub (Jun 7, 2024): This works for me!
GiteaMirror commented 2026-04-22 07:24:10 -05:00
Author
Owner
Copy Link

@vincent-guo commented on GitHub (Jun 7, 2024):

i just added this parameter OLLAMA_FLASH_ATTENTION=1 when running ollama serve, and qwen2 worked on my gpu

Well done.

<!-- gh-comment-id:2154362453 --> @vincent-guo commented on GitHub (Jun 7, 2024): > i just added this parameter OLLAMA_FLASH_ATTENTION=1 when running ollama serve, and qwen2 worked on my gpu Well done.
GiteaMirror commented 2026-04-22 07:24:10 -05:00
Author
Owner
Copy Link

@dstaicova commented on GitHub (Jun 7, 2024):

Not working for me with "ollama run". I get "GladGladGlad" (to infinity). And "ollama serve" complains "Error: accepts 0 arg(s), received 1". Any idea what is the problem?

<!-- gh-comment-id:2154527310 --> @dstaicova commented on GitHub (Jun 7, 2024): Not working for me with "ollama run". I get "GladGladGlad" (to infinity). And "ollama serve" complains "Error: accepts 0 arg(s), received 1". Any idea what is the problem?
GiteaMirror commented 2026-04-22 07:24:10 -05:00
Author
Owner
Copy Link

@Leroy-X commented on GitHub (Jun 7, 2024):

Qwen2 runs perfectly on the latest 0.1.42 update.

<!-- gh-comment-id:2154551017 --> @Leroy-X commented on GitHub (Jun 7, 2024): Qwen2 runs perfectly on the latest 0.1.42 update.
GiteaMirror commented 2026-04-22 07:24:11 -05:00
Author
Owner
Copy Link

@dstaicova commented on GitHub (Jun 7, 2024):

Yeah, well, my ollama is still 0.1.30 (manjaro) and arch's latest is ollama 0.1.41-1 as far as I can see. I guess the problem is that Manjaro doesn't prioritize ollama imports. Oh well, I'll think of a workaround.

<!-- gh-comment-id:2154574150 --> @dstaicova commented on GitHub (Jun 7, 2024): Yeah, well, my ollama is still 0.1.30 (manjaro) and arch's latest is ollama 0.1.41-1 as far as I can see. I guess the problem is that Manjaro doesn't prioritize ollama imports. Oh well, I'll think of a workaround.
GiteaMirror commented 2026-04-22 07:24:11 -05:00
Author
Owner
Copy Link

@dllp007king commented on GitHub (Jun 7, 2024):

写一个有关于windows用户的详细的解决方案
1.更新你的ollama到最新版本
2.系统找到环境变量中添加一条内容名字为OLLAMA_FLASH_ATTENTION值为1的环境变量
3.重启你的电脑
就解决了必须重启windows
1
2

<!-- gh-comment-id:2154584116 --> @dllp007king commented on GitHub (Jun 7, 2024): 写一个有关于windows用户的详细的解决方案 1.更新你的ollama到最新版本 2.系统找到环境变量中添加一条内容名字为OLLAMA_FLASH_ATTENTION值为1的环境变量 3.重启你的电脑 就解决了必须重启windows <img width="370" alt="1" src="https://github.com/ollama/ollama/assets/41323573/9eba5b8a-0480-4ea3-b27d-1d0cf5566b97"> <img width="481" alt="2" src="https://github.com/ollama/ollama/assets/41323573/76dd78a4-7669-4a94-846f-fd4c803debbc">
GiteaMirror commented 2026-04-22 07:24:12 -05:00
Author
Owner
Copy Link

@Leroy-X commented on GitHub (Jun 7, 2024):

image
0.1.42 is now a Pre-release version, available for download on the GitHub release page

<!-- gh-comment-id:2154620395 --> @Leroy-X commented on GitHub (Jun 7, 2024): ![image](https://github.com/ollama/ollama/assets/13515498/8974f14f-cb6a-4d10-9b98-e99c99916b0e) 0.1.42 is now a Pre-release version, available for download on the GitHub release page
GiteaMirror commented 2026-04-22 07:24:12 -05:00
Author
Owner
Copy Link

@jmorganca commented on GitHub (Jun 7, 2024):

Hi all sorry about this, it was fixed in Hi all, sorry about this. It is fixed in 0.1.42: https://github.com/ollama/ollama/releases/tag/v0.1.42

<!-- gh-comment-id:2155671868 --> @jmorganca commented on GitHub (Jun 7, 2024): Hi all sorry about this, it was fixed in Hi all, sorry about this. It is fixed in 0.1.42: https://github.com/ollama/ollama/releases/tag/v0.1.42
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#28846
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?