🎯Skills9
Enables Hugging Face Hub operations via the hf CLI, covering model/dataset/space downloads, file uploads, repository creation, cache management, and cloud compute job execution on HF infrastructure.
Trains and fine-tunes language models using TRL (Transformer Reinforcement Learning) on Hugging Face Jobs infrastructure, supporting SFT, DPO, GRPO, reward modeling, and GGUF conversion for local deployment.
Creates and manages datasets on Hugging Face Hub with support for SQL-based querying via DuckDB, streaming row updates, multi-format dataset support, and integration with the HF MCP server for dataset workflows.
Adds and manages evaluation results in Hugging Face model cards, supporting evaluation table extraction, Artificial Analysis API score imports, and custom model evaluations with vLLM/lighteval/inspect-ai.
Runs any workload on fully managed Hugging Face Jobs infrastructure including data processing, batch inference, experiments, and scheduled tasks on cloud CPUs/GPUs/TPUs without local setup.
Official Hugging Face skill for publishing and managing research papers on the Hub, including paper page creation, arXiv integration, model/dataset linking, authorship management, and markdown-based article generation
Tracks and visualizes ML training experiments with Trackio, providing a Python API for logging metrics during training and a CLI for retrieving metrics, with real-time dashboard visualization and HF Space syncing.
Creates reusable command-line scripts and utilities for the Hugging Face API, enabling chaining, piping, and intermediate data processing for model/dataset/space operations with proper token authentication.
Skill