HYPERPARAM OPEN-SOURCE

Summary from the TLDR newsletter

Hyperparam addresses the lack of a user-friendly, scalable UI for exploring and curating massive datasets in the machine learning ecosystem through a suite of open-source tools that tackle different aspects of data curation. It aims to empower ML practitioners to create the best training datasets for the best models. The tools include Hyparquet, a lightweight, pure-JS library for reading Apache Parquet files directly in the browser; Hyparquet-Writer, which provides a way to write or export data to Parquet format in JavaScript; HighTable, a React-based virtualized table component for viewing extremely large tables in the browser; and Icebird, which extends Hyperparam's reach into data stored in Apache Iceberg format.