Implementation of Rainbow-DQN agent for state-of-the-art discrete action design space exploration.
Execute:
conda env create --file environment.yml
conda activate rainbowOptional:
Upgrade PyTorch:
conda install pytorch torchvision torchaudio pytorch-cuda=12.1 -c pytorch -c nvidiaTested on gym-super-mario-bros environment.
- Complete comments and docstrings
- Add flexibility in environment for either 1D or 2D observations
- Enrich this README (main page) with a description of the project, results, lab page, etc.