LeRobot
Red Hat and HP Show a Self-Improving Robot Policy Loop on One AI Station
Red Hat and HP have shown a closed loop for physical AI that collects data, trains, evaluates, signs and delivers robot policies on a single HP ZGX Fury AI station.…
Anyscale Uses Ray and a Physics Twin to Probe What a Robot Policy Sees
Anyscale has described how it used Ray to run controlled experiments on a frozen vision-language-action policy inside a physics twin of its own robot. According to the…
LeRobot Now Trains on Lance Data Straight From Object Storage
LeRobot can now read Lance datasets natively, pulling frames directly from object storage with a true global shuffle and no download step. According to LanceDB, training…