How to run this?
#3
by
Autumnlight
- opened
How can I run this? I usually run exllama2 or Koboldcpp but both dont seem to support this.
Hi, We have added instructions in the model card on running the model using vLLM or transformers.
Trying to run the model from transformers - AutoModelForCausalLM
I get an authentication error - Access to model ai21labs/AI21-Jamba-1.5-Mini is restricted. You must be authenticated to access it.
Can you please provide some authentication insturctions?
@OmerAtEasily you'll need to:
- accept the model terms in the Model card
- get an Access Token from your huggingface settings page, as explained here: https://huggingface.co/docs/hub/en/security-tokens
- pass the token to
transformers(explained below)
There are several ways to let transformers use your token:
- use the
hugginface-cli loginCLI command as explained here: https://huggingface.co/docs/huggingface_hub/en/guides/cli#huggingface-cli-login. It will write the token to a file in whichtransformerswill automatically look for the token if exists - set it in
HF_TOKENenv var as explained here: https://huggingface.co/docs/huggingface_hub/en/package_reference/environment_variables#hftoken - pass the token with the
token=arg when loading the model/tokenizer as shown here: https://huggingface.co/docs/hub/en/security-tokens#how-to-use-user-access-tokens