1 paper
Omkar Patil, Ondrej Biza, Thomas Weng +9
What happens when a pretrained generative robot policy is provided a constant initial noise as input, rather than repeatedly sampling it from a Gaussian? We demonstrate that the pe…