Loading the SOTA2 catalog…
Reinforcement Learning on DMControl Ball in cup, catch (500k steps) benchmark leaderboard · SOTA2 Research