Skip to main content

HF Authentication

To access private or gated models from Hugging Face, you need to authenticate first.

Official Documentation​

Refer to the official guide for more details:
Hugging Face Authentication Guide

Step 1: Log in via CLI​

Run the following command and follow the prompts to enter your Hugging Face token:

huggingface-cli login
Don’t have a token yet?

You can create one here: https://huggingface.co/settings/tokens

Step 2: Export as ENV Variable​

To ensure your scripts or services can access the token, export it in your environment:

export HUGGING_FACE_HUB_TOKEN=<your-hf-token>

Step 3: Download a Model​

cd models/
huggingface-cli download meta-llama/Llama-3.2-1B-Instruct --local-dir Llama-3.2-1B-Instruct
LLMBoost will download it automatically.

When the specified model is not available locally, LLMBoost will automatically download it from Hugging Face, provided you specify a valid model repository (e.g. --model_name meta-llama/Llama-3.2-1B-Instruct)