[GH-ISSUE #10195] hf pull has broken #6689

Closed
opened 2026-04-12 18:25:05 -05:00 by GiteaMirror · 9 comments
Owner

Originally created by @vigsterkr on GitHub (Apr 9, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/10195

What is the issue?

when trying to pull a model from HF, which worked in the past, is currently failing with Error: unexpected status code 302. Probably this is because of the new HF storage xet

Relevant log output

ollama run https://hf.co/bartowski/deepcogito_cogito-v1-preview-qwen-14B-GGUF:Q6_K_L
pulling manifest
pulling 4bf2b2822140...   0% ▕                                                                                                                                                                                                                ▏    0 B/ 12 GB
Error: unexpected status code 302

OS

Docker

GPU

No response

CPU

No response

Ollama version

No response

Originally created by @vigsterkr on GitHub (Apr 9, 2025). Original GitHub issue: https://github.com/ollama/ollama/issues/10195 ### What is the issue? when trying to pull a model from HF, which worked in the past, is currently failing with `Error: unexpected status code 302`. Probably this is because of the new HF storage `xet` ### Relevant log output ```shell ollama run https://hf.co/bartowski/deepcogito_cogito-v1-preview-qwen-14B-GGUF:Q6_K_L pulling manifest pulling 4bf2b2822140... 0% ▕ ▏ 0 B/ 12 GB Error: unexpected status code 302 ``` ### OS Docker ### GPU _No response_ ### CPU _No response_ ### Ollama version _No response_
GiteaMirror added the bug label 2026-04-12 18:25:05 -05:00
Author
Owner

@nviraj commented on GitHub (Apr 9, 2025):

Facing this as well on Windows 11, latest version of ollama (v0.6.5)

<!-- gh-comment-id:2789799841 --> @nviraj commented on GitHub (Apr 9, 2025): Facing this as well on Windows 11, latest version of ollama (v0.6.5)
Author
Owner

@rick-github commented on GitHub (Apr 9, 2025):

Until #10197 is merged, if you have spare disk space you can create your own q6 quant:

ollama pull cogito:14b-v1-preview-qwen-fp16
echo FROM cogito:14b-v1-preview-qwen-fp16 > Modelfile
ollama create -q q6_k cogito:14b-v1-preview-qwen-q6_k
<!-- gh-comment-id:2789913785 --> @rick-github commented on GitHub (Apr 9, 2025): Until #10197 is merged, if you have spare disk space you can create your own q6 quant: ```console ollama pull cogito:14b-v1-preview-qwen-fp16 echo FROM cogito:14b-v1-preview-qwen-fp16 > Modelfile ollama create -q q6_k cogito:14b-v1-preview-qwen-q6_k ```
Author
Owner

@rick-github commented on GitHub (Apr 9, 2025):

The ollama team has rejected the fix, so pulling this model (and perhaps others) requires a change on the HF side.

<!-- gh-comment-id:2790614069 --> @rick-github commented on GitHub (Apr 9, 2025): The ollama team has rejected the fix, so pulling this model (and perhaps others) requires a change on the HF side.
Author
Owner

@rick-github commented on GitHub (Apr 9, 2025):

Pulling models from HuggingFace is unsupported (#10205), it just happens to mostly work at the moment.

<!-- gh-comment-id:2790850373 --> @rick-github commented on GitHub (Apr 9, 2025): Pulling models from HuggingFace is unsupported (#10205), it just happens to mostly work at the moment.
Author
Owner

@julien-c commented on GitHub (Apr 10, 2025):

We'll take a look! cc @xcid @vaibhavs10

We'll try to fix on HF side but we might need to open a PR on ollama cc @jmorganca

<!-- gh-comment-id:2792314016 --> @julien-c commented on GitHub (Apr 10, 2025): We'll take a look! cc @xcid @vaibhavs10 We'll try to fix on HF side but we might need to open a PR on ollama cc @jmorganca
Author
Owner

@n00b001 commented on GitHub (Apr 10, 2025):

FYI while this is not fixed, if people need to run their models somewhere you could use:
https://github.com/aphrodite-engine/aphrodite-engine

It has HuggingFace integration that can successfully download from huggingface

<!-- gh-comment-id:2793292768 --> @n00b001 commented on GitHub (Apr 10, 2025): FYI while this is not fixed, if people need to run their models somewhere you could use: https://github.com/aphrodite-engine/aphrodite-engine It has HuggingFace integration that can successfully download from huggingface
Author
Owner

@XciD commented on GitHub (Apr 10, 2025):

ollama run https://hf.co/bartowski/deepcogito_cogito-v1-preview-qwen-14B-GGUF:Q6_K_L

is fixed

<!-- gh-comment-id:2794494947 --> @XciD commented on GitHub (Apr 10, 2025): ``` ollama run https://hf.co/bartowski/deepcogito_cogito-v1-preview-qwen-14B-GGUF:Q6_K_L ``` is fixed
Author
Owner

@bmizerany commented on GitHub (Apr 10, 2025):

We'll take a look! cc @XciD @Vaibhavs10

We'll try to fix on HF side but we might need to open a PR on ollama cc @jmorganca

@julien-c You can ping me. I handle 99.9% of the registry work and API.

<!-- gh-comment-id:2794601960 --> @bmizerany commented on GitHub (Apr 10, 2025): > We'll take a look! cc [@XciD](https://github.com/XciD) [@Vaibhavs10](https://github.com/Vaibhavs10) > > We'll try to fix on HF side but we might need to open a PR on ollama cc [@jmorganca](https://github.com/jmorganca) @julien-c You can ping me. I handle 99.9% of the registry work and API.
Author
Owner

@bmizerany commented on GitHub (Apr 10, 2025):

Confirmed this issue is with the external registry and not Ollama. Closing.

<!-- gh-comment-id:2794615569 --> @bmizerany commented on GitHub (Apr 10, 2025): Confirmed this issue is with the external registry and not Ollama. Closing.
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/ollama#6689