Fetching the paper…

Action Image Representation: Learning Scalable Deep Grasping Policies with Zero Real World Data · Around