Use this inside Hugging face API

#4
by joncc - opened

Newbie question - how do I load the tokenizer and ".pt" model into a Huggingface pipeline using the HF API?

I've been trying stuff like "

model = LlamaForCausalLM.from_pretrained("./alpaca7b-4bit.pt")

and

tokenizer = LlamaTokenizer.from_pretrained("./tokenizer.model")

But no success

This is a very old quant at this point, I don't recommend using it. Take a look at TheBloke's work isntead.

ozcur changed discussion status to closed

Thanks for the reply

Sign up or log in to comment