ResearchBenchmarksRL Motion Imitation on Unitree H1-2 Pistol SquatFollow383.7Final RewardDDR193.38242.79292.2341.61May 22, 2026Evaluation ResultsMethodMethodLinksFinal RewardSteps to 90% RewardDDRReference Type=Direct...Reference Type=Direct Dynamic Retargeting2026.05383.779.6IDRReference Type=Immedia...Reference Type=Immediate Physics Correction2026.05340.893.5GRReference Type=Kinemat...Reference Type=Kinematic Retargeting2026.05200.7128.4