Continuous Control on MuJoCo HalfCheetah v4 (test)
15,227Mean Episodic ReturnFP-DRL
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| FP-DRL2026.04 | 15,227 | 512 | |
| SAC2026.04 | 13,629 | 994 | |
| SAC-FLOW2026.04 | 12,609 | 0 | |
| DSAC-T2026.04 | 11,694 | 171 | |
| DACER2026.04 | 10,277 | 399 | |
| TD32026.04 | 8,824 | 0 | |
| RATArchitecture=Shared actor-critic networks, Seeds=5 random seeds2026.05 | 4,629.2 | — | |
| PPOArchitecture=Shared actor-critic networks, Seeds=5 random seeds2026.05 | 4,146 | — | |
| ACKTRArchitecture=Shared actor-critic networks, Seeds=5 random seeds2026.05 | 3,630.9 | — | |
| SophiaArchitecture=Shared actor-critic networks, Seeds=5 random seeds2026.05 | 899.5 | — |