Continuous Control on Walker2d v5
6,138.2Avg ReturnDBC
Evaluation results
| Method | Links | ||
|---|---|---|---|
| DBCActor Backbone=SAC2026.02 | 6,138.2 | — | |
| TRFP(ours)NFE=4 × 42026.04 | 5,986.9 | — | |
| TRFP(one-step)NFE=1 × 42026.04 | 5,933.9 | — | |
| TQCActor Backbone=SAC2026.02 | 5,802.6 | — | |
| SACNFE=12026.04 | 5,665.4 | — | |
| TD3NFE=12026.04 | 4,883.4 | — | |
| SACActor Backbone=SAC2026.02 | 4,854.4 | — | |
| IQNActor Backbone=SAC2026.02 | 4,766.5 | — | |
| SDACNFE=20 × 322026.04 | 3,906.5 | — | |
| PPOEnvironment steps=1M, Seeds=10, Tests per epoch=102026.03 | 3,796.7 | — | |
| PDAEnvironment steps=1M, Seeds=10, Tests per epoch=102026.03 | 3,642.3 | — | |
| MaxEntDPNFE=20 × 102026.04 | 3,628.5 | — | |
| DSACActor Backbone=SAC2026.02 | 3,297.4 | — | |
| VFActor Backbone=SAC2026.02 | 2,626.4 | — | |
| VDActor Backbone=SAC2026.02 | 2,533.9 | — | |
| RPIDDPG2025.06 | 319.9 | 245 | |
| DDPG2025.06 | 200.3 | 165.7 |