Download Model
If the model is already on the NAS
Skip this step and proceed to 04. Register Model.
In the Jupyter Lab terminal, download a Hugging Face model and save it to a Volume.
Run in the terminal, not a notebook cell
Open a terminal via the Jupyter top menu File > New > Terminal and run there.
If you enter it into a notebook cell (.ipynb) Python interprets it and you'll get a SyntaxError.
Install required library
pip install -U huggingface_hub
Download a public model (example: Qwen2.5-0.5B-Instruct)
hf download Qwen/Qwen2.5-0.5B-Instruct \
--local-dir /data/Qwen2.5-0.5B-Instruct
Storage path
Set --local-dir to the mount path configured when creating the Volume. The example above assumes the mount path is /data.
Saving outside the Volume (e.g., /tmp) loses the data when the Lab Pod terminates.
Download a private model (requires a HuggingFace token)
hf login --token $HF_TOKEN
hf download <org>/<model> \
--local-dir /data/<model-name>
Check the download
ls /data/Qwen2.5-0.5B-Instruct
The download is healthy if you see the following files:
config.json
tokenizer.json
tokenizer_config.json
*.safetensors