ByteDance builds TableVerse, a 100,000-scene dataset for tabletop robot manipulation
ByteDance · hf · 2026-07-24
ByteDance’s TableVerse builds 100,000 tabletop scenes from real images
ByteDance presents TableVerse, a fully automated Real2Sim pipeline for generalizable robotic manipulation.
Instead of generating layouts from text or simplified procedural rules, TableVerse reconstructs simulation-ready tabletop environments from unstructured in-the-wild images. The pipeline aims to preserve:
- metric scale
- authentic topology
- mechanical stability
It also adds an automated task-conditioned trajectory generator to synthesize collision-free pick-and-place demonstrations. Using this pipeline, the team built TableVerse-100K, a dataset of 100,000 unique environments paired with manipulation trajectories, intended as a scalable foundation for future manipulation research.
More from Embodied
- Tencent's HiSTEVE 1.0: hierarchical VLA agent plays Minecraft with native keyboard and mouse, hits 85% SOTA — _akhaliq · 2026-10-26
- SJTU's VaM lets robots learn online from mistakes, 95%+ success over 4-hour autonomous runs — chris_j_paxton · 2026-10-26
- Inside Dexory's full-stack warehouse robotics factory: 8m scanning towers and in-house manufacturing — lukas_m_ziegler · 2026-10-26
- M5Stack StickS3 firmware adds remote live view for Ricoh GR III/IIIx/IV cameras — OpenAIDevs · 2026-10-26
- Fourier's GR-3 humanoid climbs six floors autonomously, descends stairs untethered — chris_j_paxton · 2026-10-26
- FocusVLA routes robot camera views by predicted interaction intent across task stages — iscas-cloud · 2026-10-26