hf

Log in to Hugging Face Hub

$ hf auth login

$ hf download [gpt2]

$ hf download [meta-llama/Llama-2-7b] [config.json] [model.safetensors]

$ hf upload [username/my-model] [./models] [.]

$ hf models ls

$ hf jobs run --flavor [a10g-small] [pytorch/pytorch:2.6.0-cuda12.4-cudnn9-devel] [python train.py]

$ hf cache ls

hf command [subcommand] [options]

hf (formerly huggingface-cli) is the official command-line interface for Hugging Face Hub. It provides direct terminal access to download models and datasets, upload files, manage repositories, run compute jobs, and interact with the ML community platform.
The CLI follows a consistent hf resource action pattern. Files are cached locally in ~/.cache/huggingface/hub/ with smart deduplication. The download command supports partial downloads with --include/--exclude patterns and resumable transfers.
hf jobs enables running code on Hugging Face infrastructure including GPUs (T4, A10G, A100) and TPUs, with Docker-like commands. Jobs support environment variables, secrets, scheduled execution, and UV scripts for self-contained Python tasks.
Authentication is managed via hf auth login which stores tokens locally. Tokens can also be set via the HF_TOKEN environment variable for scripting.

--repo-type type

Repository type: model, dataset, or space.

Specific revision (branch, tag, or commit hash).

Download to specific local directory instead of cache.

Include files matching glob pattern.

Exclude files matching glob pattern.

Authentication token for private repos.

Suppress verbose output, print only final result.

Hardware for jobs: cpu-basic, t4-small, a10g-small, a100-large, etc.

Job timeout with units: 30m, 2h, 1d.

Display help for any command.

auth login

Log in to Hugging Face Hub with access token.

Log out and remove stored tokens.

Display current user and organizations.

Download files from the Hub to local cache.

Upload files or folders to a repository.

Create a new repository on the Hub.

Delete a repository.

List cached repositories and their sizes.

Remove cached files.

Run compute jobs on Hugging Face infrastructure.

List and search models on the Hub.

List and search datasets.

List and search Spaces.

Print environment information for debugging.

Some features require a Hugging Face Pro subscription or organization membership, including hf jobs for compute access. Large model downloads may require significant disk space and bandwidth. Private repositories require authentication.

The Hugging Face Hub CLI was introduced as part of the huggingface_hub Python package to provide programmatic access to the model repository. In July 2025, the CLI was renamed from huggingface-cli to hf for brevity, with commands reorganized into a clearer resource action structure. The legacy command remains functional with deprecation warnings.

ollama(1), git-lfs(1), docker(1)