LEADERBOARD

OBSTACLES, DISCRETE
OBSTACLES, CONTINUOUS

The best result in OffWorldMonolithContinuousReal-v0 environment is 0.88 by Felix Lu!
Run your policy in evalutaion mode by calling  gym.make(..., mode='test')  to see it here.

Experiments conducted: 96
Episodes finished: 22092
Hours of real learning logged: 552

TOP 3 IN END TO END


SAC-REAL-Continuous-7
by Felix Lu, average test reward 0.88

SAC-REAL-Continuous-6
by Felix Lu, average test reward 0.87

Macro_net6
by Karl Aru, average test reward 0.80

TOP 3 IN SIM TO REAL


TOP 3 IN HUMAN DEMONSTRATIONS