API Key rejected?

@bogdog400 It should be that token. It should be starting with "hf_" if its a correct token.

I can't really replicate this, in worst case could you mail me an example token that I could try to replicate it with? You could mail on huggingface@marcusmailbox.com. As soon as I have tested it you can revoke the token.

I was trying to use some of the meta-llama/Meta-Llama models (and some other ones) and also got the same error, and other warnings. In the end I got it working with these steps:

1. Create a Token
2. Update permissions under "Access Tokens > YOUR_TOKEN > Manage > Edit Permissions" and set these values:
   

   Inference
   x Make calls to the serverless Inference API
   x Make calls to Inference Endpoints
     Manage Inference Endpoints
   Repos
   x Read access to contents of all public gated repos you can access
   [...]
   

3. I got the error The model meta-llama/Meta-Llama-3-8B is too large to be loaded automatically, after changing to meta-llama/Meta-Llama-3-8B-Instruct it worked

See The Serverless Inference API: "The model meta-llama/Meta-Llama-3-8B is too large to be loaded automatically (16GB > 10GB)".