Huggingface
A comprehensive skill catalog for AI agents
npx -y skills add G1Joshi/Agent-Skills --skill huggingfaceAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
One thing to look at
- 10 stars10 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
What its author says it does
Copied from the file, not written here
Hugging Face transformers library and hub. Use for NLP models.
SKILL.md
1.2 KB, as published. Nobody here has run it
Hugging Face
Hugging Face is the GitHub of AI. It hosts 1M+ models. 2025 sees massive growth in Multimodal models and Robotics (LeRobot).
When to Use
- Model Discovery: Finding the SOTA open-source model for any task.
- Inference:
transformerslibrary is the standard way to run models in Python. - Datasets: Accessing standard datasets (
load_dataset('squad')).
Core Concepts
Transformers Library
The API to download and run models. pipeline('sentiment-analysis').
Hugging Face Hub (Hugging Face CLI)
Versioning, git-based storage for large model weights (git lfs).
Spaces
Hosting simple Gradio/Streamlit apps for model demos.
Best Practices (2025)
Do:
- Use
bitsandbytes: Load 70B models in 4-bit precision easily. - Use
accelerate: For multi-GPU training/inference distributed across devices. - Push to Hub: Share your fine-tunes.
Don't:
- Don't hardcode paths: Use
from_pretrained("repo/id")to auto-cache models.