hf-download

Retry Hugging Face downloads automatically with resume and lock cleanup.

3|2|Updated Mar 14, 2026
One-click install
npx skills add https://github.com/wzh4464/claude-skills --skill hf-download
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: hf-download
Source: https://github.com/wzh4464/claude-skills/tree/main/hf-download
Command: npx skills add https://github.com/wzh4464/claude-skills --skill hf-download

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Large Hugging Face downloads can time out, fail mid-transfer, or stall due to stale lock files; this skill provides an auto-retry wrapper that transparently resumes failed downloads and cleans up locks.

Core Features & Use Cases

  • Auto-retry wrapper for hf download that handles timeouts, connection drops, and stale lock files.
  • Resumes interrupted transfers and uses token-based rate limiting when needed.
  • Works with a variety of Hugging Face models and can be invoked as a Claude skill or directly via the hf-download.sh script.
  • Supports extra hf download flags and environment variable configurations for token and transfer settings.

Quick Start

Run the hf-download.sh script with a repository ID and optional local directory and flags to start auto-retrying downloads of Hugging Face models.

Frequently Asked Questions about hf-download

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

FAQPage Schema
Why do Hugging Face model downloads time out and how can I resume them?

Hugging Face downloads time out due to flaky networks or stale lock files. You can resume interrupted transfers by using an auto-retry wrapper that transparently cleans up locks and restarts failed fetches.

How do I automatically retry failed hf download commands for large LLMs?

You can automatically retry failed hf download commands by running the hf-download.sh script with a repository ID, optional local directory, and extra flags to enable auto-retrying transfers for large LLMs.

Can I pass extra flags and tokens to Hugging Face downloads for rate limiting?

Yes, you can pass extra flags and environment variables to configure token-based rate limiting. This helps manage transfer settings and avoid hitting rate limits during large model fetches.

What is the best way to fix stale lock files during Hugging Face model fetches?

The best way to fix stale lock files during Hugging Face model fetches is to use an auto-retry wrapper that transparently cleans up stale locks before resuming the interrupted download transfer.

Does the auto-retry Hugging Face download wrapper work with diffusion models?

Yes, the auto-retry Hugging Face download wrapper works with diffusion models and other sizable assets. It reliably retrieves large files over flaky networks by resuming dropped connections.