Paper
Robótica & RL
HumanScale: Egocentric Human Video Can Outperform Real-Robot Data for Embodied Pretraining
Egocentric human video can effectively replace teleoperated robot trajectories for embodied model pretraining, achieving better performance with reduced data collection costs.
Hugging Face · Daily Papers
·Juncheng Ma, Jianxin Bi
·
·▲ 8 upvotes
Este artigo está em destaque na seleção diária de papers do Hugging Face, curada pela comunidade de pesquisa em IA.
Autores: Juncheng Ma, Jianxin Bi, Yufan Deng, Xuanran Zhai, Kewei Zhang, Ye Huang
- 8 upvotes da comunidade
- Temas: embodied foundation models, teleoperated real-robot trajectories, egocentric human video, pretraining, action prediction, task execution
Resumo
Resumo original (em inglês), extraído do paper:
Egocentric human video can effectively replace teleoperated robot trajectories for embodied model pretraining, achieving better performance with reduced data collection costs.
// relacionados
