It provides a PyQt6 desktop app that lets users load data, apply LLM‑derived business constraints, train TabICL classifiers or regressors, view metrics, and forecast time‑series. The tool includes offline bundling scripts and integrates with a local LM Studio server for knowledge injection. It targets data scientists and ML engineers who want a user‑friendly way to leverage TabICL without writing code. Compared to command‑line usage, it streamlines the workflow and adds optional LLM‑based data validation.
View on GitHub →yoannfradin-sudo/Tabcl-Gui