huggingface-cli

Login to Hugging Face Hub

$ huggingface-cli login

$ huggingface-cli whoami

$ huggingface-cli logout

$ huggingface-cli env

$ huggingface-cli download --repo-type [repo_type] [repo_id] [filename1 filename2 ...]

$ huggingface-cli upload --repo-type [repo_type] [repo_id] [path/to/local_file_or_directory] [path/to/repo_file_or_directory]

$ huggingface-cli scan-cache

$ huggingface-cli delete-cache

huggingface-cli [-h] [--version] [-v] command [args]
where command is one of: login, logout, whoami, scan-cache, delete-cache, env, repo

login
    Log in to Hugging Face Hub using token or browser

logout
    Log out and clear authentication token

whoami
    Display current authenticated user info

scan-cache
    Scan and list local cache contents

delete-cache
    Delete specific cache entries by repo or pattern

env
    Print relevant environment variables

repo
    Subcommands for repo management (create, ls, info, etc.)

--token
    HF token for authentication (global for some cmds)

-v, --verbose
    Increase output verbosity

-h, --help
    Show help for command or subcommand

--version
    Display CLI version

The huggingface-cli is the official command-line interface for interacting with the Hugging Face Hub, a platform hosting thousands of machine learning models, datasets, and applications.

It enables users to authenticate, manage repositories, scan and clean local caches, and perform various Hub operations without a graphical interface. Common tasks include logging in with an access token, creating or listing repos, checking user info, and optimizing disk space by deleting cached files.

Designed for developers and researchers, it integrates seamlessly with Python workflows via the huggingface_hub library. Requires Python 3.8+ and installation via pip install -U huggingface_hub. Ideal for CI/CD pipelines, server environments, and local machine maintenance.

Requires valid HF account and token for private repos. Cache paths depend on HF_HOME env var. Not all Git LFS features supported directly.

pip install -U huggingface_hub
Tokens from https://huggingface.co/settings/tokens

huggingface-cli login
huggingface-cli repo create my-model
huggingface-cli scan-cache --verbose

Introduced in huggingface_hub v0.4.0 (2022) as lightweight CLI wrapper. Evolved with Hub features; v0.20+ added repo subcommands and improved cache mgmt.

git(1), pip(1), curl(1)