MAP670C Reinforcement Learning - Course assignments
In order to run this project here are few steps you have to complete
Step 0 : Open your favorite terminal
Step 1 : Clone the repo
$ git clone https://github.com/Minozar/rl-laboratory
Step 2 : Create a python virtual environment
$ cd rl-laboratory
$ python -m venv venv-name
Step 3 : Run the virtual environment
# Windows PowerShell
$ venv-name\Scripts\activate.ps1
# Linux terminal (bash)
$ source venv-name/bin/activate
Step 4 : Install the dependencies
$ pip install -r requirements.txt
Now you are ready to go !
The project is quite simple at the moment To start the benchmarking, simply run the following command :
$ python ./main.py
And then, to compute the final graph :
$ python ./benchmark_results.py
Our first Soft Actor Critic best model trained on BipedalWalker-v3 with total_timesteps = 500000.
Our first Double Deterministic Policy Gradient best model trained on BipedalWalker-v3 with total_timesteps = 500000.
Our first Proximal Policy Optimization best model trained on BipedalWalker-v3 with total_timesteps = 500000.
