Training used 27,300 hours of real-robot and egocentric video.
Chris Paxton#732
Excited to release τ0-WM: an open-source unified video-action world model for robotic manipulation.
It's a 5B-parameter robotic foundation model trained on 27.3K hours of real-robot teleoperation, UMI-style demonstrations, and egocentric interaction videos.
Robotics is so fascinating
They talk about scaling, and then train a 5B model on 30k hours of data
Excited to release τ0-WM: an open-source unified video-action world model for robotic manipulation.
It's a 5B-parameter robotic foundation model trained on 27.3K hours of real-robot teleoperation, UMI-style demonstrations, and egocentric interaction videos.
Training used 27,300 hours of real-robot and egocentric video.
Chris Paxton#732Excited to release τ0-WM: an open-source unified video-action world model for robotic manipulation.
It's a 5B-parameter robotic foundation model trained on 27.3K hours of real-robot teleoperation, UMI-style demonstrations, and egocentric interaction videos.
Some users like the idea of learning while deploying in the open-source 5B τ0-WM video-action world model for robot manipulation.