π0.7 training · Kuan Yi Wang
acastapaloma.github.io · saved by 2 readers
An illustrated walkthrough of π0.7: how images, text, and joint positions become robot actions, how the model is trained, and which models write its prompt.
An illustrated walkthrough of π0.7: how images, text, and joint positions become robot actions, how the model is trained, and which models write its prompt.