Let Offline RL Flow: Training Conservative Agents in the Latent Space of Normalizing Flows

Open in new window