Google DeepMind RT-2Google DeepMind RT-2Google DeepMind RT-2

Vision-language-action (VLA) model by Google DeepMind for robotic control, combining web-scale vision-language pretraining with robotics data, announced July 28, 2023.