D4RL
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
D4RL Antmaze-umaze-diverse 1M (full)
71.4Normalized Score
4
D4RL Locomotion Stochastic
59.4Hopper (Med-Expert)
4
D4RL AntMaze Large Diverse Offline → Online
85Offline Success Rate
3
D4RL AntMaze Large Play Offline → Online
86Success Rate (Offline)
3
D4RL AntMaze Medium Diverse Offline → Online
76Offline Success Rate
3
D4RL AntMaze Medium Play Offline → Online
77Success Rate (Offline)
3
D4RL AntMaze U-Maze Diverse Offline → Online
93Success Rate (Offline)
3
D4RL AntMaze U-Maze Offline → Online
100Success Rate (Offline)
3
D4RL Hopper-expert
113.08Score (Normalized)
3
D4RL HalfCheetah expert
107.74Normalized Score (100k steps)
3
D4RL
11.52Offline Training Time (hour)
3
D4RL
19.96Offline Training Time
3
D4RL
20.96Training Time (hour)
3
D4RL HalfCheetah-random
0.807RMAE
3
D4RL antmaze-large-diverse v2
94D4RL Score
2
D4RL antmaze-large-play v2
93D4RL Score
2
D4RL antmaze-medium-diverse v2
96.5D4RL Score
2
D4RL antmaze-medium-play v2
89.5D4RL Score
2
D4RL antmaze-umaze-diverse v2
99.8D4RL Score
2
D4RL antmaze-umaze v2
99.5D4RL Score
2
D4RL maze2d-large v1
220.7D4RL Score
2
D4RL maze2d-medium v1
166.8D4RL Score
2
D4RL maze2d-umaze v1
161.6D4RL Score
2
D4RL Hopper-medium-expert
0.192RMAE
2
D4RL Hopper-medium
0.208RMAE
2