[privategpt] HuggingFace Token

[mdima]

Hello,
I am trying to start the container, but I receive the error:

Downloading tokenizer mistralai/Mistral-7B-Instruct-v0.2
Traceback (most recent call last):
  File "/home/worker/app/.venv/lib/python3.11/site-packages/huggingface_hub/utils/_errors.py", line 270, in hf_raise_for_status
    response.raise_for_status()
  File "/home/worker/app/.venv/lib/python3.11/site-packages/requests/models.py", line 1021, in raise_for_status
    raise HTTPError(http_error_msg, response=self)
requests.exceptions.HTTPError: 401 Client Error: Unauthorized for url: https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.2/resolve/main/config.json

So I read here zylon-ai/private-gpt#1861 that a Token is required to download that file.

I added, as suggested above, the env variable:
--env=HUGGINGFACE_TOKEN=xxxxxxxxxxxxx
and I accepted the agreement for the model "Mistral-7B-Instruct-v0.2" in HuggingFace, but I receive the same error.

Can you please help?

Thanks,
Michele

[3x3cut0r]

hey,
the problem is, that mistralai now wants you to agree to the terms. this is also an upstream issue.
anyways ... i did not implement the env HUGGINGFACE_TOKEN either.
now i added it and uploaded a new image.
try now to provide your own token and it should work.

but I still have to see how this is solved at the upstream private-gpt repo as well

[mdima]

Thank you very much! I confirm it works now...
Of course, better if this issue will be solved in the upstream repository, but for now the issue is solve.

Thanks again,
Michele