github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-08 00:51:34 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #8045] Ollama run hf.co - Error 401: Invalid username or password #51655

New Issue
Closed
opened 2026-04-28 20:42:27 -05:00 by GiteaMirror · 9 comments
GiteaMirror commented 2026-04-28 20:42:27 -05:00
Owner
Copy Link

Originally created by @bengrau on GitHub (Dec 11, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/8045

What is the issue?

I am using a private model on hf and try to run it like this:

huggingface-cli login --token hf_xxx
ollama run hf.co/BGR/Llama-3.2-1B-I-p:latest

However I get this error from ollama:

pulling manifest 
Error: pull model manifest: 401: {"error":"Invalid username or password."}

Anyone experienced this issue or knows how to solve this?

OS

Linux

GPU

Nvidia

CPU

AMD

Ollama version

ollama version is 0.5.1

Originally created by @bengrau on GitHub (Dec 11, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/8045 ### What is the issue? I am using a private model on hf and try to run it like this: ``` huggingface-cli login --token hf_xxx ollama run hf.co/BGR/Llama-3.2-1B-I-p:latest ``` However I get this error from ollama: ``` pulling manifest Error: pull model manifest: 401: {"error":"Invalid username or password."} ``` Anyone experienced this issue or knows how to solve this? ### OS Linux ### GPU Nvidia ### CPU AMD ### Ollama version ollama version is 0.5.1
GiteaMirror added the bug label 2026-04-28 20:42:27 -05:00
GiteaMirror commented 2026-04-28 20:42:30 -05:00
Author
Owner
Copy Link

@zaidalyafeai commented on GitHub (Dec 20, 2024):

Hey @bengrau, have you solved this? I want to access a gated model on HuggingFace.

<!-- gh-comment-id:2556750583 --> @zaidalyafeai commented on GitHub (Dec 20, 2024): Hey @bengrau, have you solved this? I want to access a gated model on HuggingFace.
GiteaMirror commented 2026-04-28 20:42:30 -05:00
Author
Owner
Copy Link

@pdevine commented on GitHub (Dec 20, 2024):

I'm going to close this since there's not really any solution on the Ollama side for this. I'm not sure how HuggingFace implemented that; Ollama has a particular way that it does access control (with ed25519 public/private keys) which I'm guessing they haven't implemented.

As a workaround, you can pull the gguf file directly and create a Modelfile with the line:

FROM Llama-3.2-1B-I-p.gguf # or whatever the file is called

and then running:

ollama create my-model
<!-- gh-comment-id:2557784338 --> @pdevine commented on GitHub (Dec 20, 2024): I'm going to close this since there's not really any solution on the Ollama side for this. I'm not sure how HuggingFace implemented that; Ollama has a particular way that it does access control (with ed25519 public/private keys) which I'm guessing they haven't implemented. As a workaround, you can pull the gguf file directly and create a Modelfile with the line: ``` FROM Llama-3.2-1B-I-p.gguf # or whatever the file is called ``` and then running: ``` ollama create my-model ```
GiteaMirror commented 2026-04-28 20:42:30 -05:00
Author
Owner
Copy Link

@bengrau commented on GitHub (Dec 21, 2024):

Couldn't find a solution and did it the way @pdevine described it.

<!-- gh-comment-id:2558000625 --> @bengrau commented on GitHub (Dec 21, 2024): Couldn't find a solution and did it the way @pdevine described it.
GiteaMirror commented 2026-04-28 20:42:32 -05:00
Author
Owner
Copy Link

@michetonu commented on GitHub (Jan 2, 2025):

Had the same problem, turns out you need to add your Ollama SSH key to your Huggingface account. You can follow the instructions on this page:

You can run private GGUFs from your personal account or from an associated organisation account in two simple steps:

    - Copy your Ollama SSH key, you can do so via: cat ~/.ollama/id_ed25519.pub | pbcopy
    - Add the corresponding key to your Hugging Face account by going to [your account settings](https://huggingface.co/settings/keys) and clicking on Add new SSH key.
    - That’s it! You can now run private GGUFs from the Hugging Face Hub: ollama run hf.co/{username}/{repository}.
<!-- gh-comment-id:2567852366 --> @michetonu commented on GitHub (Jan 2, 2025): Had the same problem, turns out you need to add your Ollama SSH key to your Huggingface account. You can follow the instructions on [this page](https://huggingface.co/docs/hub/en/ollama#run-private-ggufs-from-the-hugging-face-hub): ``` You can run private GGUFs from your personal account or from an associated organisation account in two simple steps: - Copy your Ollama SSH key, you can do so via: cat ~/.ollama/id_ed25519.pub | pbcopy - Add the corresponding key to your Hugging Face account by going to [your account settings](https://huggingface.co/settings/keys) and clicking on Add new SSH key. - That’s it! You can now run private GGUFs from the Hugging Face Hub: ollama run hf.co/{username}/{repository}. ```
GiteaMirror commented 2026-04-28 20:42:32 -05:00
Author
Owner
Copy Link

@Master-Pr0grammer commented on GitHub (Apr 4, 2025):

I did what @michetonu suggested but im still getting the same error. Trying to download the new quantized aware training gemma 3 4b model google released, but keep getting this error.

Edit: as a temporary solution, someone downloaded the gemma 3 QAT models and re-uploaded it here: https://www.reddit.com/r/LocalLLaMA/comments/1jqyfs9/ollama_fix_gemma312bitqatq4_0gguf/

<!-- gh-comment-id:2778555198 --> @Master-Pr0grammer commented on GitHub (Apr 4, 2025): I did what @michetonu suggested but im still getting the same error. Trying to download the new quantized aware training gemma 3 4b model google released, but keep getting this error. Edit: as a temporary solution, someone downloaded the gemma 3 QAT models and re-uploaded it here: https://www.reddit.com/r/LocalLLaMA/comments/1jqyfs9/ollama_fix_gemma312bitqatq4_0gguf/
GiteaMirror commented 2026-04-28 20:42:33 -05:00
Author
Owner
Copy Link

@kkishore9891 commented on GitHub (Apr 5, 2025):

I did what @michetonu suggested but im still getting the same error. Trying to download the new quantized aware training gemma 3 4b model google released, but keep getting this error.

Same. The pbcopy command is apparently a MacOS command. We Linux users are ffed.

<!-- gh-comment-id:2780866805 --> @kkishore9891 commented on GitHub (Apr 5, 2025): > I did what [@michetonu](https://github.com/michetonu) suggested but im still getting the same error. Trying to download the new quantized aware training gemma 3 4b model google released, but keep getting this error. Same. The pbcopy command is apparently a MacOS command. We Linux users are ffed.
GiteaMirror commented 2026-04-28 20:42:33 -05:00
Author
Owner
Copy Link

@ed7coyne commented on GitHub (Apr 19, 2025):

Just copy the output of that cat and paste it into the box on the huggingface UI. This is similar to the github workflow for ssh keys.

<!-- gh-comment-id:2816782514 --> @ed7coyne commented on GitHub (Apr 19, 2025): Just copy the output of that cat and paste it into the box on the huggingface UI. This is similar to the github workflow for ssh keys.
GiteaMirror commented 2026-04-28 20:42:34 -05:00
Author
Owner
Copy Link

@JamesClarke7283 commented on GitHub (Apr 23, 2025):

I'm going to close this since there's not really any solution on the Ollama side for this. I'm not sure how HuggingFace implemented that; Ollama has a particular way that it does access control (with ed25519 public/private keys) which I'm guessing they haven't implemented.

As a workaround, you can pull the gguf file directly and create a Modelfile with the line:

FROM Llama-3.2-1B-I-p.gguf # or whatever the file is called

and then running:

ollama create my-model

It wont work, it doesnt convert the chat template metadata, i tried with glm4-Z1-9b, it just gave the following template in the model file:

TEMPLATE {{ .Prompt }}

that model was a reasoning one, so even if it worked somewhat, it wouldnt have reasoning as it needed something in the template to activate it.

<!-- gh-comment-id:2825428063 --> @JamesClarke7283 commented on GitHub (Apr 23, 2025): > I'm going to close this since there's not really any solution on the Ollama side for this. I'm not sure how HuggingFace implemented that; Ollama has a particular way that it does access control (with ed25519 public/private keys) which I'm guessing they haven't implemented. > > As a workaround, you can pull the gguf file directly and create a Modelfile with the line: > > ``` > FROM Llama-3.2-1B-I-p.gguf # or whatever the file is called > ``` > > and then running: > > ``` > ollama create my-model > ``` It wont work, it doesnt convert the chat template metadata, i tried with glm4-Z1-9b, it just gave the following template in the model file: ``` TEMPLATE {{ .Prompt }} ``` that model was a reasoning one, so even if it worked somewhat, it wouldnt have reasoning as it needed something in the template to activate it.
GiteaMirror commented 2026-04-28 20:42:34 -05:00
Author
Owner
Copy Link

@JamesClarke7283 commented on GitHub (Apr 23, 2025):

I did what @michetonu suggested but im still getting the same error. Trying to download the new quantized aware training gemma 3 4b model google released, but keep getting this error.

Same. The pbcopy command is apparently a MacOS command. We Linux users are ffed.

Lol, good one.

Regardless it raises a good point so i will adapt the command for GNU/Linux and while i am at it some other platforms, see below for the guide.

Most free/libre POSIX Systems (Linux, BSD, Hurd (: )

xclip (X11)

cat ~/.ollama/id_ed25519.pub | xclip -sel c

wl-clipboard (Wayland)

cat ~/.ollama/id_ed25519.pub | wl-copy

Windows

Windows Terminal

type %USERPROFILE%\.ollama\id_ed25519.pub | clip

Powershell

Get-Content $env:USERPROFILE\.ollama\id_ed25519.pub | Set-Clipboard

For Copilot+PC/Recall

Recall users have a shortcut.

If you haven't disabled Recall, it might have a copy stored (if you ever opened it). You could ask Copilot to retrieve the pub-key and put it in your clipboard, or anything else ever displayed on-screen... and that's not frightening at all!

Hope someone finds this guide useful.

Good day ;)

Image

Cropped Version of 'This is fine' by KC Green of GunShowComic

<!-- gh-comment-id:2825516129 --> @JamesClarke7283 commented on GitHub (Apr 23, 2025): > > I did what [@michetonu](https://github.com/michetonu) suggested but im still getting the same error. Trying to download the new quantized aware training gemma 3 4b model google released, but keep getting this error. > > Same. The pbcopy command is apparently a MacOS command. We Linux users are ffed. Lol, good one. Regardless it raises a good point so i will adapt the command for GNU/Linux and while i am at it some other platforms, see below for the guide. ## Most free/libre POSIX Systems (Linux, BSD, Hurd (: ) ## xclip (X11) ```sh cat ~/.ollama/id_ed25519.pub | xclip -sel c ``` ## wl-clipboard (Wayland) ```sh cat ~/.ollama/id_ed25519.pub | wl-copy ``` ## Windows ### Windows Terminal ```batch type %USERPROFILE%\.ollama\id_ed25519.pub | clip ``` ### Powershell ```powershell Get-Content $env:USERPROFILE\.ollama\id_ed25519.pub | Set-Clipboard ``` ### For Copilot+PC/Recall Recall users have a shortcut. If you haven't disabled Recall, it might have a copy stored (if you ever opened it). You could ask Copilot to retrieve the pub-key and put it in your clipboard, or anything else ever displayed on-screen... and that's not frightening at all! --- Hope someone finds this guide useful. Good day ;) --- <img src="https://github.com/user-attachments/assets/bdb019e1-678f-4ca1-b06b-b0376fc778a1" alt="Image" alt="'This is fine' by KC Green" width="512" height="288"> ### Cropped Version of **_['This is fine'](https://gunshowcomic.com/648) by [KC Green](https://kcgreendotcom.com/) of [GunShowComic](https://gunshowcomic.com/)_**
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#51655
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?