1d agoSystems engineer Yacine trains a stable RL policy in under three seconds using Pufferlib and MuJoCo WarpRL speed is now constrained by environment simulation speed.SentimentSentimentPos100%Neg0%Many users praise Pufferlib's Cartpole training speeds in MuJoCo because Nvidia Warp enables incredible performance that they describe as slick and solid.5 comments with sentiment. View comments.