Hugging Face Dataset Viewer

0•

Explores, filters, and extracts data from Hugging Face datasets using the Dataset Viewer API and SQL queries.

The Hugging Face Dataset Viewer skill provides a robust interface for Claude to interact with the Hugging Face Hub, allowing for seamless dataset exploration and data extraction. It enables users to validate dataset availability, resolve configurations and splits, and preview or paginate through rows using the official API. Beyond simple browsing, the skill supports advanced features like full-text search, predicate-based filtering, and the ability to execute SQL queries on Parquet shards via parquetlens. It is an essential tool for data scientists and AI researchers who need to inspect, audit, or prepare machine learning datasets directly within their development workflow.

主な機能

01Row-level pagination and previews with configurable offsets
02SQL query execution on remote Parquet shards via parquetlens
030 GitHub stars
04Metadata retrieval for dataset subsets and splits
05Dataset statistics and Croissant metadata extraction
06Advanced text search and predicate-based row filtering

ユースケース

01Inspecting dataset samples and structure before starting model training
02Extracting specific data subsets using SQL and filtering predicates
03Automating the validation and exploration of newly uploaded Hugging Face datasets