A plain vanilla Deep Q Network with multi-layer perceptron for any control tasks. The model we are going to test is Cartpole-v0
you would be requiring the following
-
Python 3.6 or above click here
-
Jupyter notebook/ lab
python3 -m pip install --upgrade pip
python3 -m pip install jupyter
jupyter notebook
- Tensorflow click for installation
- Numpy & matplotlib
pip install numpy
pip install matplotlib
- Deep Q Network (plain)
- Double DQN
- Duel
- Noisy exploration
- Distributional
- Fixed Q Targets
- Prioritized Experience Replay