RAD_PPO, RAD_SAC are not included in this repo now?

ac-93 / tactile_gym

Suite of PyBullet reinforcement learning environments targeted towards using tactile data as the main form of observation.

GNU General Public License v3.0

127 stars 20 forks source link

RAD_PPO, RAD_SAC are not included in this repo now? #2

Closed emigmo closed 3 years ago

emigmo commented 3 years ago

The sb3_contrib package can not import the RAD_SAC and RAD_PPO, And current sb3_contrib repo can only support TQC and QR-DQN. So we can not reproduce the performance now?

ac-93 commented 3 years ago

I've a fork of sb3_contrib here that I used for training, still slowly working on reproducing results to contribute to sb3_contrib.

You can clone that repo and install locally in order to use RAD_PPO and RAD_SAC. With that said, standard PPO will for the most part produce similar performance, augmentations just tended to make training more stable in my experience.