Content of error returned by a model not yet authorized is not accurate

[edgararuiz]

I was testing v3, which I haven't authorized yet on the PriorLabs website, and it returned:

  HuggingFace authentication error downloading from 'Prior-Labs/tabpfn_3'. This model is gated and requires you to accept its terms.  Please
  follow these steps: 1. Visit https://huggingface.co/Prior-Labs/tabpfn_3 in your browser and accept the terms of use. 2. Log in to your
  Hugging Face account via the command line by running: hf auth login (Alternatively, you can set the HF_TOKEN environment variable with a
  read token.) For detailed instructions, see https://docs.priorlabs.ai/how-to-access-gated-models

The link https://huggingface.co/Prior-Labs/tabpfn_3 is invalid, takes me to a 404 page in HuggingFace. Also, I believe going to HuggingFace is not the way to authorize, but we all need to go to the PriorLabs portal for pre-authorization

It is not technically a bug, or even a doc request, but the issue here is that new folks are gonna be confused about where to go to get authorized

TabPFN/src/tabpfn/errors.py

Lines 59 to 72 in 4b4bdaa

	def __init__(self, repo_id: str):
	message = (
	f"HuggingFace authentication error downloading from '{repo_id}'.\n"
	"This model is gated and requires you to accept its terms.\n\n"
	"Please follow these steps:\n"
	f"1. Visit https://huggingface.co/{repo_id} in your browser and"
	f" accept the terms of use.\n"
	"2. Log in to your Hugging Face account via"
	" the command line by running:\n"
	" hf auth login\n"
	" (Alternatively, you can set the HF_TOKEN environment variable"
	" with a read token.)\n\n"
	"For detailed instructions, see "
	"https://docs.priorlabs.ai/how-to-access-gated-models"

[naga-karthik]

Facing the same issue, the HF link works now but still getting HF authentication error (even though hf auth login is done and model was downloaded from HF Hub)