Offline-to-Online Reinforcement Learning on D4RL Gym-Locomotion benchmark leaderboard · SOTA2 Research

Offline-to-Online Reinforcement Learning on D4RL Gym-Locomotion

88.36HalfCheetah Return (Random)ROAD

Evaluation Results

Method	Links
ROAD2026.05		88.36	77.67	86.94	92.1	96.86	32.08	53.5	63.63	79.04	79.63	34.46	68.43	58.61	70.81	57.58
Decreasing Mixing Strategy2026.05		87.35	62.57	78.86	89.68	93.86	15.93	34.47	48.84	67.07	52.22	23.01	49.1	39.33	50.2	64.74
Fixed Mixing Ratio2026.05		80.43	73.44	82.71	89.44	93.75	16.54	34.83	50.49	41.67	37.11	13.49	42.6	35.98	50.27	40.26
Uniform Mixing Strategy2026.05		80.07	71.28	83.53	88.14	93.17	16.82	35.95	40.51	66.29	48.57	22.67	52.17	42.27	39.78	60.24
Fixed Mixing Ratio2026.05		79.11	64.26	73.39	88.35	89.55	12.06	34.85	38.86	70.61	49.26	36.61	45.73	39.23	58.79	43.52
Fixed Mixing Ratio2026.05		78.16	72.68	83.27	89.9	93.61	15.77	38.89	40.42	66.36	37.64	21.65	40.81	33.57	49.53	41.82
Fixed Mixing Ratio2026.05		78.07	78.73	81.44	88.2	93.48	27.14	36.08	39.66	57.74	36.99	29.82	38.44	40.29	42.45	58.57
BR2026.05		76.94	72.51	83.37	91.08	96.28	31.63	42.21	37.47	23.03	89.06	17.81	47.3	41.43	48.83	49.84
Fixed Mixing Ratio2026.05		76.65	60.32	78.48	87.35	90.73	13.84	28.39	34.01	53.73	50.31	25.2	55.99	34.07	56.38	43.06
Fixed Mixing Ratio2026.05		75.82	74.2	85.3	89.28	94.44	28.2	36.24	41.55	61.69	46.23	20.79	55.97	34.55	66.13	62.31