The post gives the model's scale and joint video and action prediction design, which helps readers see how video pretraining connects to robot control.
Part of a thread · earlier post
Praxis-1 is a 27B-parameter mixture-of-experts diffusion transformer that predicts video and actions together. We detail the model's architecture, training stages and results in our technical report.
https://runway.com/research/praxis-1-video-pretraining-to-robot-control
Runway · x.comPublished · added here
Original title: Praxis-1 is a 27B-parameter mixture-of-experts diffusion transformer that predicts video and actions together. We detail the model's arch...
