TD3+BC, D4RL: Evaluation for Best Hyperparameters

Tables 2, 3, 4 in the paper

Created on May 15|Last edited on May 15

Comment

﻿
HalfCheetah﻿
eval/normalized_score_mean
eval/normalized_score_mean
2004006008001k1.2kStep020406080
dataset_name: halfcheetah-full-replay-v2, actor_bc_coef: 0.01, critic_bc_coef: 0
dataset_name: halfcheetah-expert-v2, actor_bc_coef: 0.4, critic_bc_coef: 0
dataset_name: halfcheetah-random-v2, actor_bc_coef: 0.001, critic_bc_coef: 0
dataset_name: halfcheetah-medium-replay-v2, actor_bc_coef: 0.05, critic_bc_coef: 0
dataset_name: halfcheetah-medium-expert-v2, actor_bc_coef: 0.1, critic_bc_coef: 0
dataset_name: halfcheetah-medium-v2, actor_bc_coef: 0.01, critic_bc_coef: 0
Run set8358
﻿
Hopper﻿
﻿
Run set5923
﻿
Walker2d﻿
Run set8358
﻿
AntMaze﻿
Run set8358
﻿
Pen﻿
Run set8358
﻿
Door﻿
Run set8358
﻿
Hammer﻿
Run set8358
﻿
Relocate﻿
Run set8358
﻿
﻿

Add a comment