README.md

🧱 LEGO Brick Classifier — Gradio Space

This Space wraps a trained AutoGluon tabular model that classifies LEGO pieces as Standard, Flat, or Sloped from four inputs: Length, Height, Width, Studs.

• Model repo: Iris314/classical-automl-model (uses lego_predictor_dir.zip)
• Dataset: aedupuga/lego-sizes
• Model type: AutoML ensemble (best = LightGBM)
• Purpose: Coursework demo for classical ML + Gradio

How it works

• Downloads the predictor from the model repo via huggingface_hub.
• Renames UI features to the training schema: Length → Max Length (cm), Height → Max Height (cm), Width → Width (cm), keeps Studs.
• Runs predict_proba and shows a top-3 probability label.

Notes

• This is a teaching demo; dataset is small/synthetic. Not for production.