Rohan Paul (@rohanpaul_ai)
2025-05-11 | โค๏ธ 525 | ๐ 76
LegoGPT, an LLM-based system that generates physically stable LEGO structures from text prompts, backed by a new 47,000+ sample dataset and physics-aware filtering during inference.
โ LegoGPT is trained on a custom dataset, StableText2Lego, which includes 47,000+ 3D LEGO models mapped to text, spanning 28,000+ unique objects.
โ The model predicts LEGO bricks sequentially like tokens, using next-token prediction in a transformer setup.
โ To ensure physical stability, LegoGPT integrates physics-aware rollback and validity filtering, pruning out structurally invalid brick placements.
โ The generated designs are aesthetically aligned with prompts, physically buildable, and tested both with human manual assembly and robotic arms.
โ The team also introduced a text-driven LEGO coloring/texturing pipeline, enabling more expressive and customized outputs.
โ The dataset, code, and models are all publicly released under an open-access license.
๋ฏธ๋์ด
๐ Related
Auto-generated - needs manual review
Tags
domain-vision-3d domain-robotics domain-ai-ml domain-simulation domain-dev-tools domain-crypto domain-visionos