Skill UI

Weights & Biases (W&B) automates experiment logging, real-time dashboards, hyperparameter sweeps, artifact tracking, and model registry/versioning for ML teams using PyTorch, TensorFlow, or HuggingFace workflows.

This professional tool helps developers and researchers identify the optimal AI model for a specific task, such as coding, reasoning, or image recognition. It queries official HuggingFace leaderboards, filters models based on hardware constraints (e.g., available VRAM or RAM), and provides a detailed comparison table including benchmark scores, parameter size, and licensing information.

A comprehensive guide and skill for deploying large language models (LLMs) locally using the llama.cpp framework and GGUF quantization format. It covers searching the Hugging Face Hub, selecting optimal quantizations (e.g., Q4_K_M), running local servers, and achieving OpenAI-compatible API endpoints. Supports various hardware backends including CPU, Mac Metal, CUDA, and ROCm. Ideal for developers needing offline, private LLM inference.

Automates the creation and publication of polished, functional Gradio demos on Hugging Face Spaces for any user-provided LoRA model. This skill handles the complex workflow of designing the optimal UI, selecting the correct inference pipeline (e.g., text-to-image, video-to-video), and generating all necessary files (app.py, requirements.txt, README.md) to deploy a complete, ready-to-use AI playground.

This skill guides users through building, deploying, and maintaining machine learning applications on Hugging Face Spaces. It provides comprehensive knowledge on selecting the optimal SDK (Gradio, Docker, Static), choosing appropriate hardware resources (ZeroGPU, dedicated GPUs, CPU-basic), and following best practices to ensure ML applications are deployed efficiently and successfully on the platform.

A comprehensive utility for building reusable, command-line scripts and utilities to interact with the Hugging Face API. It enables complex data workflows by allowing users to chain, pipe, and process data fetched from models, datasets, and various endpoints. Ideal for automation, enrichment, and advanced data pipelines.

A comprehensive guide for developers building AI demos and GPU compute tasks on Hugging Face Spaces using ZeroGPU hardware. It covers best practices for writing and reviewing code that utilizes the `@spaces.GPU` decorator, managing runtime duration and quota, and understanding the specific CUDA availability model and process isolation rules required for reliable ML deployment.

This skill automates the collection of high-quality technical news from multiple sources like Hacker News and HuggingFace. It uses a multi-agent architecture to scrape, filter, and deduplicate content, generating a curated daily Markdown report. Ideal for staying updated on cutting-edge technology and AI developments without manual searching.

This skill helps users find the best AI models for specific tasks by querying Hugging Face benchmark leaderboards. It analyzes hardware constraints like GPU VRAM or CPU memory to filter compatible models. The output includes a comparison table with benchmark scores, parameter sizes, licenses, and device compatibility status. It also suggests whether models can run locally or require API access.

This skill guides users on running evaluations against models on the Hugging Face Hub using local hardware. It covers using inspect-ai and lighteval frameworks with local inference backends like vllm, Transformers, or accelerate. It includes task selection, smoke testing, and hardware guidance for different model sizes. It does not cover remote orchestration or publishing results to the community workflow.

Use this skill for Hugging Face Dataset Viewer API workflows that fetch subset/split metadata, paginate rows, search text, apply filters, download parquet URLs, and read size or statistics. It also covers creating and uploading datasets, and uploading agent traces.

This skill helps you select and run large language models locally using llama.cpp and GGUF formats. It covers finding compatible models on the Hugging Face Hub, choosing quantization levels for different hardware like CPU, Mac Metal, CUDA, or ROCm, and launching servers via llama-cli or llama-server. It also includes instructions for converting weights and verifying OpenAI-compatible local serving endpoints.