Question about convert-hf-to-gguf-update.py
When starting it I need to enter . Does this mean I need to enter my Access Tokens?
Because I did and when trying to download llama-bpe I get an error - Failed to download file. Status code: 403
I was able to download the files, only it seems to happen after I accessed the Meta repository. But I'm not sure.
Yes you use ...-update.py <hf-read-token>.
Some files will give you a 403 error - they don't exist in the remote - but the ones we need should be downloaded fine inside models\tokenizers\llama-bpe.
Replace the files in the model folder with the ones in the llama-bpe folder.
The thing is that I was able to download them only when I got access to Meta-Llama-3. That is, before I did not gain access to their repository, I was given a 403 error - that is, access was denied.
I just don’t know whether this should be included in the README or not, that to download these files you need access to the Meta repository.
Huh, I think this should be included. I never thought about that. Lemme see if I have access to them...
Yeah, indeed. Oh well then!
Because, for example, I cannot get these files from the token of this account. But from the second account, on which I was able to access Meta-Llama-3, I can get these files.
Added to the Warning. All the people I talked about so far I think just were too into it and already had access, haha. Thanks for the reminder.
Not sure if I would get in trouble for hosting these myself...
Let it be just a warning. There's nothing complicated about it.
It's inside the [!WARNING] that was already there, just added more context about the repos access to fetch the files. Should be alright.