M2T2

M2T2 (Multi-Task Masked Transformer) is a unified transformer model for learning different primitive actions.

Primary Use Cases

Given a raw point cloud observation of the scene, M2T2 reasons about contact points and predicts collision-free gripper poses for 6-DoF object-centric grasping and orientation-aware placement.

Date

This model was trained in June 2023.

Resources for More Information

• Paper: https://arxiv.org/abs/2311.00926
• Code: https://github.com/NVlabs/M2T2
• Website: https://m2-t2.github.io

Citation

If you find our work helpful, please consider citing our paper.

@inproceedings{yuan2023m2t2,
  title     = {M2T2: Multi-Task Masked Transformer for Object-centric Pick and Place},
  author    = {Yuan, Wentao and Murali, Adithyavairavan and Mousavian, Arsalan and Fox, Dieter},
  booktitle = {7th Annual Conference on Robot Learning},
  year      = {2023}
}