Couldn't reproduce the result on Mujoco suite.

[sweetice]

Couldn't reproduce the result on the Mujoco suite.
Setting: We run the BEAR with the recommend settings: ** mmd_sigma = 20.0 , kernel_type = gaussian , num_samples_match = 5 , version = 0 or 2 , lagrange_thresh = 10.0 , `mode = auto**
The batch dataset is produced by training a DDPG agent for 1 million time steps. For reproducing, we use the DDPG code in BCQ repository.

We utilize the final buffer setting in BCQ paper.
Here are the whole results.
Note that the "behavioral" means the evaluation of the DDPG agent when training.

Sweetice 0720 update: Upload the png file for easily reading.

For more clear reading.
Ant-v2.pdf
HalfCheetah-v2.pdf
Hopper-v2.pdf
InvertedDoublePendulum-v2.pdf
InvertedPendulum-v2.pdf
Reacher-v2.pdf
Swimmer-v2.pdf
Walker2d-v2.pdf

[aviralkumar2907]

Hi! In the BEAR paper, we didn't test on the final buffer setting, so I am not sure if the hyperparameters are optimal for this setting as well. I would also recommend trying out the new cleaned up implementation of BEAR, where we have searched over hyperparameters: https://github.com/rail-berkeley/d4rl_evaluations and would also recommend using the D4RL datasets (https://github.com/rail-berkeley/d4rl).