RoboticsRobotics: Science and Systems (RSS)2023年10月

RT-1: Robotics Transformer for Real-World Control at Scale

Anthony Brohan, Noah Brown, Justice Carmona, Johnny Conchou, Krista Choromanski, Chelsea Finn, Danny Driess, Aviral Kumar, Abby Harrison, Hanjun KimGoogle Research

摘要

We present RT-1, a Robotics Transformer trained on a large-scale real-world robotics dataset. Our model demonstrates that transformers can effectively learn from diverse real-world robot manipulation data and generalize to novel tasks and objects in unseen environments.

robotics transformerreal-world controlmanipulationscalable learningembodied AI

技术细节

测试机器人
模型骨架

Transformer

编码器

EfficientNet

解码器

Token Learner

相关公司(2)

京ICP备2026064258号-1