Hugging Face Skills

Uses the Hugging Face Hub CLI to download, upload and manage models, datasets, Spaces, buckets and jobs.

Runs evaluations of Hugging Face Hub models on local hardware with inspect-ai and lighteval.

Works with the Hugging Face Dataset Viewer API: split metadata, row pagination, text search, filters and parquet downloads.

Builds Gradio web UIs and demos in Python: components, event listeners, layouts and chatbots.

Fine-tunes language and vision models with TRL or Unsloth on Hugging Face Jobs, including GGUF conversion.

Helps select and run models locally with llama.cpp and GGUF on CPU, Mac Metal, CUDA or ROCm.

Publishes and manages research papers on the Hugging Face Hub and links them to models and datasets.

Reads Hugging Face paper pages as markdown and fetches structured metadata such as authors and linked models.

Builds, deploys and maintains apps on Hugging Face Spaces with the Gradio, Docker or Static SDKs.

Tracks and visualises ML training experiments with Trackio: logged metrics, alerts and a live dashboard.

Trains and fine-tunes vision models for object detection, image classification and SAM segmentation.

Runs machine learning models directly in JavaScript and TypeScript with Transformers.js, for text, vision and audio tasks.