What problem does it solve? Interacting with the Hugging Face Hub through the web UI or raw API calls is slow and hard to automate. This Skill provides a complete reference for the modern hf command-line interface, replacing the deprecated huggingface-cli, so you can script and automate Hub operations directly from the terminal. ## Core Features & Use Cases - Model & Dataset Operations: Search, download, and upload models and datasets, including resumable uploads of large folders and SQL queries against dataset parquet files via DuckDB. - Repository & Collaboration Management: Create, move, duplicate, and delete repositories, manage branches and tags, and handle discussions and pull requests including diffs and merges. - Infrastructure & Automation: Deploy Inference Endpoints, run compute jobs, manage Spaces, S3-like buckets, webhooks, and local cache. - Use Case: A machine learning engineer needs to publish a fine-tuned model. They authenticate with hf auth login, create a repository with hf repos create, upload weights with hf upload-large-folder, and tag a release version, all from the command line. ## Quick Start Use the huggingface-hub skill to download the model meta-llama/Llama-3-8B from the Hugging Face Hub to my local machine.
