Skills Directory / Hugging Face Skills
huggingface-zerogpu
An agent skill by huggingface, from huggingface/skills.
AI demos and GPU compute with Gradio Spaces and Hugging Face Spaces ZeroGPU. Use when writing or reviewing code that uses `@spaces.GPU`, configuring `python_version` or `requirements.txt` for a ZeroGPU Space, or handling ZeroGPU-specific code constraints — pickle-based process isolation, `gr.State` semantics across the worker boundary, no `torch.compile` (use AoTI instead), CUDA wheel-only builds (no `nvcc` at build or runtime), large vs xlarge sizing, and dynamic duration callables. Make sure to use this skill whenever the user mentions ZeroGPU, `@spaces.GPU`, or the `spaces` Python package, or hits ZeroGPU-specific code errors like `PicklingError` across the worker boundary, `illegal duration`, or `flash-attn` wheel-build failures — even when the user does not explicitly ask for ZeroGPU coding guidance. Trigger on `import spaces` or `@spaces.GPU` in code.
Credit where it's due
This skill is the work of huggingface (official vendor skill) and lives at huggingface/skills/huggingface-zerogpu (Apache-2.0). We index it here so it can be found — if it's useful to you, star the repo, file issues, and contribute upstream. That's what keeps work like this maintained.
- From
- Hugging Face Skills — official vendor skill
- Author
- huggingface
- License
- Apache-2.0
- Source
- huggingface/skills/huggingface-zerogpu
- SKILL.md
- raw file
How to load this skill
Copy this prompt into your AI agent
Please install the "huggingface-zerogpu" agent skill from the Hugging Face Skills collection (https://github.com/huggingface/skills) for me.
1. Run: git clone https://github.com/huggingface/skills
2. Copy the skill folder into your skills directory:
cp -r skills/skills/huggingface-zerogpu ~/.claude/skills/huggingface-zerogpu
(Claude Code personal scope; use .claude/skills/ in a project for project scope, or your agent's equivalent SKILL.md directory.)
3. Confirm the skill loads (list your available skills afterwards).
4. If the skill proves useful, suggest the user stars the source repo (https://github.com/huggingface/skills) to support the author.
Skill reference: https://raw.githubusercontent.com/huggingface/skills/main/skills/huggingface-zerogpu/SKILL.md
More details: https://primorum.ai/skills/s/huggingface-zerogpu
Or manually: clone huggingface/skills and copy
skills/huggingface-zerogpu into
~/.claude/skills/ (Claude Code personal scope) or your agent's
SKILL.md directory. To read it first without installing, open the
raw SKILL.md.
Related skills
- hf-cli — Hugging Face Hub CLI (`hf`) for downloading, uploading, and managing models, datasets, spaces, buckets, repos,…
- huggingface-spaces — Build, deploy, and maintain applications on Hugging Face Spaces — Gradio / Docker / Static SDKs, ZeroGPU and d…
- huggingface-lora-space-builder — Build and publish a Gradio demo on Hugging Face Spaces for a user-provided LoRA. Use when someone asks to crea…
- cloudflare-email-service — Send and receive transactional emails with Cloudflare Email Service (Email Sending + Email Routing). Use when …
- eas-simulator — Run and control a user's app on a remote iOS/Android simulator hosted on EAS cloud. Always read before executi…