github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-07 08:30:05 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #1881] Only generate lots of hashes #1076

New Issue
Closed
opened 2026-04-12 10:49:47 -05:00 by GiteaMirror · 9 comments
GiteaMirror commented 2026-04-12 10:49:47 -05:00
Owner
Copy Link

Originally created by @ZhihaoZhang97 on GitHub (Jan 10, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/1881

Screenshot from 2024-01-10 11-52-07

Not sure if I am the first to encounter with this issue, when I installed the ollama and run the llama2 from the Quickstart, it only outputs a lots of '####'.

I suspect that might be caused by the hardware or software settings with my newly updated system?

Since it works with my old rig with i9-9900K and dual RTX 3090.

As shown in the screenshot below, I am currently using Pop!OS with AMD Threadripper 3960X and dual RTX 3090.
Screenshot from 2024-01-10 11-52-54

Any help would be greatly appreciated, thank you!

Originally created by @ZhihaoZhang97 on GitHub (Jan 10, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/1881 ![Screenshot from 2024-01-10 11-52-07](https://github.com/jmorganca/ollama/assets/31653817/30f08c0d-c924-471f-b740-896ba804c2bf) Not sure if I am the first to encounter with this issue, when I installed the ollama and run the llama2 from the Quickstart, it only outputs a lots of '####'. I suspect that might be caused by the hardware or software settings with my newly updated system? Since it works with my old rig with i9-9900K and dual RTX 3090. As shown in the screenshot below, I am currently using Pop!OS with AMD Threadripper 3960X and dual RTX 3090. ![Screenshot from 2024-01-10 11-52-54](https://github.com/jmorganca/ollama/assets/31653817/ebc410c8-d635-4d7c-9d31-9115d67b1516) Any help would be greatly appreciated, thank you!
GiteaMirror commented 2026-04-12 10:49:49 -05:00
Author
Owner
Copy Link

@lasseedfast commented on GitHub (Jan 10, 2024):

Before there has been a workaround for this, but the problem seems to be back again. Here are some more info https://github.com/jmorganca/ollama/pull/1261#issuecomment-1881823438

<!-- gh-comment-id:1884433321 --> @lasseedfast commented on GitHub (Jan 10, 2024): Before there has been a workaround for this, but the problem seems to be back again. Here are some more info https://github.com/jmorganca/ollama/pull/1261#issuecomment-1881823438
GiteaMirror commented 2026-04-12 10:49:49 -05:00
Author
Owner
Copy Link

@maciejmajek commented on GitHub (Jan 10, 2024):

Same here. Tested on: v0.1.19, v0.1.17 and docker

2x4090, i9-13900k, ubuntu 20.04
Driver Version: 545.23.08
CUDA Version: 12.1

I was able to run the models using latest version just fine for some time but at some point every output became a stream of hashes.

Edit:
mixtral outputs hashes only
phi outputs empty lines
mistral works fine

<!-- gh-comment-id:1884681869 --> @maciejmajek commented on GitHub (Jan 10, 2024): Same here. Tested on: [v0.1.19](https://github.com/jmorganca/ollama/releases/tag/v0.1.19), [v0.1.17](https://github.com/jmorganca/ollama/releases/tag/v0.1.17) and [docker](https://hub.docker.com/r/ollama/ollama) 2x4090, i9-13900k, ubuntu 20.04 Driver Version: 545.23.08 CUDA Version: 12.1 I was able to run the models using latest version just fine for some time but at some point every output became a stream of hashes. Edit: mixtral outputs hashes only phi outputs empty lines mistral works fine
GiteaMirror commented 2026-04-12 10:49:50 -05:00
Author
Owner
Copy Link

@elven2016 commented on GitHub (Jan 12, 2024):

the same error too, have you found the solution?

<!-- gh-comment-id:1888290493 --> @elven2016 commented on GitHub (Jan 12, 2024): the same error too, have you found the solution?
GiteaMirror commented 2026-04-12 10:49:51 -05:00
Author
Owner
Copy Link

@lasseedfast commented on GitHub (Jan 16, 2024):

My solution has become to downgrade to .17

curl https://ollama.ai/install.sh | sed 's#https://ollama.ai/download#https://github.com/jmorganca/ollama/releases/download/v0.1.17#' | sh.

<!-- gh-comment-id:1893224659 --> @lasseedfast commented on GitHub (Jan 16, 2024): My solution has become to downgrade to .17 ```curl https://ollama.ai/install.sh | sed 's#https://ollama.ai/download#https://github.com/jmorganca/ollama/releases/download/v0.1.17#' | sh.```
GiteaMirror commented 2026-04-12 10:49:52 -05:00
Author
Owner
Copy Link

@ZhihaoZhang97 commented on GitHub (Jan 22, 2024):

It seems downgrade the Nvidia Driver back to 535.x.x can also resolve the problem with the latest ollama.

<!-- gh-comment-id:1904968929 --> @ZhihaoZhang97 commented on GitHub (Jan 22, 2024): It seems downgrade the Nvidia Driver back to 535.x.x can also resolve the problem with the latest ollama.
GiteaMirror commented 2026-04-12 10:49:53 -05:00
Author
Owner
Copy Link

@lasseedfast commented on GitHub (Jan 23, 2024):

Thanks. If you know some up to date instructions on how to downgrade please share, I've not found any easy enough for me to follow.

<!-- gh-comment-id:1905427419 --> @lasseedfast commented on GitHub (Jan 23, 2024): Thanks. If you know some up to date instructions on how to downgrade please share, I've not found any easy enough for me to follow.
GiteaMirror commented 2026-04-12 10:49:53 -05:00
Author
Owner
Copy Link

@maciejmajek commented on GitHub (Jan 24, 2024):

Still happening. v0.1.20 + nvidia 545
Tested both locally and inside docker with and without gpus.
image
Models ran using cpu only docker image run fine.

<!-- gh-comment-id:1908255329 --> @maciejmajek commented on GitHub (Jan 24, 2024): Still happening. v0.1.20 + nvidia 545 Tested both locally and inside docker with and without gpus. ![image](https://github.com/ollama/ollama/assets/46171033/7052c561-01d5-4610-b1a9-f3813123aace) Models ran using cpu only docker image run fine.
GiteaMirror commented 2026-04-12 10:49:54 -05:00
Author
Owner
Copy Link

@pdevine commented on GitHub (Jan 27, 2024):

Sorry guys, can you try again w/ 0.1.22 and make sure you the model you're trying to use.

<!-- gh-comment-id:1912885624 --> @pdevine commented on GitHub (Jan 27, 2024): Sorry guys, can you try again w/ `0.1.22` and make sure you the model you're trying to use.
GiteaMirror commented 2026-04-12 10:49:54 -05:00
Author
Owner
Copy Link

@ZhihaoZhang97 commented on GitHub (Jan 27, 2024):

Thanks @pdevine can confirm the 0.1.22 version fix the bug with the latest Nvidia 545 driver! Nice work!
Screenshot from 2024-01-27 13-45-13

<!-- gh-comment-id:1912927953 --> @ZhihaoZhang97 commented on GitHub (Jan 27, 2024): Thanks @pdevine can confirm the `0.1.22` version fix the bug with the latest Nvidia 545 driver! Nice work! ![Screenshot from 2024-01-27 13-45-13](https://github.com/ollama/ollama/assets/31653817/6dfa17d8-430c-4243-bb87-a435af7237e1)
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
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#1076
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?