SynCity 3000: Text-to-3D Generation for Large-Scale Coherent Scenes

The SynCity 3000 paper introduces a scene-level 3D diffusion method that fine-tunes image-to-3D generators to produce large-scale, coherent 3D worlds directly from text prompts with controllable layouts.

2026-07-07 ~ 2026-07-07 · 2 related posts

• SynCity 3000: Text-to-3D Large-Scale World Generation — anselm · 2026-07-07
• SynCity 3000: Scene-Level 3D Diffusion for Large-Scale Generation — Oxford · 2026-07-07