DreamerPro: Reconstruction-Free Model-Based Reinforcement Learning with Prototypical Representations

Open in new window