Minari

The 'Minari' benchmark is a suite containing 25 continuous control tasks used to evaluate the performance of reinforcement learning algorithms.

Papers using Minari (3)

SOPE: Stabilizing Off-Policy Evaluation for Online RL with Prior Data2026Agile Reinforcement Learning through Separable Neural Architecture2026Guided Flow Policy: Learning from High-Value Actions in Offline Reinforcement Learning2025