Consistency models offer efficient and expressive policy representation for reinforcement learning, outperforming diffusion models in online RL settings.
Consistency models offer efficient and expressive policy representation for reinforcement learning, outperforming diffusion models in computational efficiency while maintaining comparable performance.