huggingface-hub

Manage models, datasets, and Spaces on the Hugging Face Hub via CLI.

247|22|Updated Dec 11, 2024
One-click install
npx skills add https://github.com/graniet/kheish --skill huggingface-hub-graniet
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: huggingface-hub
Source: https://github.com/graniet/kheish/tree/main/skills/mlops/huggingface-hub
Command: npx skills add https://github.com/graniet/kheish --skill huggingface-hub-graniet

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Manually managing machine learning assets like models, datasets, and Spaces on the Hugging Face Hub is tedious, error-prone, and difficult to scale for teams or automated MLOps pipelines, especially when working with large files or recurring workflows.

Core Features & Use Cases

  • Full Hub CLI Access: Interact with every part of the Hugging Face Hub, including searching, downloading, and uploading models and datasets, managing repositories, running SQL queries on dataset parquet files, deploying inference endpoints, and managing Spaces and cloud storage buckets.
  • Automated Workflow Support: Streamline repetitive tasks like bulk file transfers, dataset filtering, and infrastructure management without manual web UI interaction.
  • Use Case Example: A machine learning engineer can use this skill to quickly pull a pre-trained model, query a large public dataset for specific training samples, and push processed results to a private team repository in minutes, rather than hours of manual work.

Quick Start

Use the huggingface-hub skill to download the 'bert-base-uncased' model to your local environment and upload a sample inference script to your private Hugging Face Hub Space.

Frequently Asked Questions about huggingface-hub

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I automate Hugging Face Hub model and dataset management for MLOps?

Automating Hugging Face Hub management involves using the Hub CLI to script repetitive MLOps tasks. This includes programmatically searching, downloading, and uploading models or datasets, alongside deploying inference endpoints, to eliminate manual web UI interaction.

Can I use the Hugging Face Hub CLI to run SQL queries on dataset parquet files?

Running SQL queries on dataset parquet files is supported through the Hugging Face Hub CLI. This allows you to filter large public datasets for specific training samples and streamline dataset curation without downloading entire repositories locally.

Do I need an HF_TOKEN to upload models and deploy inference endpoints?

An HF_TOKEN is required for write operations like uploading models and deploying inference endpoints. You must install the official Hugging Face Hub CLI and authenticate with a valid token to enable automated repository management and team collaboration.

What is the best way to push processed machine learning results to a private Hugging Face Space?

Pushing processed machine learning results to a private Space uses the Hub CLI for direct repository uploads. This automates file transfers to your private Hugging Face Hub repository, allowing teams to share assets efficiently without manual web uploads.

Does bulk file transfer for Hugging Face Hub repositories work with large files?

Bulk file transfer for Hugging Face Hub repositories supports large files within automated workflows. By leveraging the Hub CLI, you can execute bulk transfers and manage cloud storage buckets programmatically, avoiding the errors common in manual uploads.