The code in 'generateScores.py' generates scores reported in the paper.
For Atari only 100k data models are provided because of the size limit.
Because of the same reason 2.5k data is not provided for control tasks for baselines

SAC.py contains the implementation of the SAC algorithm.
PPO.py contains the implementation of the PPO algorithm.
Trained models are the ones in the folder.

Requirements listed under requirements.txt should be installed first.

To generate scores, run the following command:

```
python generateScores.py
```

The scores will be saved in the 'scores.txt' file.

Arguments:
--turns : number of turns to run the simulation for, default is 1
--log-file: name of the log file to save the simulation scores, default is 'scores.txt'
--log-directory: directory to save the log file, default is '.'
--models-directory: directory to of the models, default is '.'