Fetching the paper…

Coarse-to-Fine Q-attention: Efficient Learning for Visual Robotic Manipulation via Discretisation · Around