Task Agnostic Dynamics Priors for Deep Reinforcement Learning

Code for Task Agnostic Dynamics Priors for Deep Reinforcement Learning

Requirements

pip install -r requirements.txt

Running

You can use python train_batch.py to run various different RL experiments of Physics environments and Atari with different flags. Pretrained models from a dataset are stored in physmodel.