Skip to content

Implementation of a DQN agent to solve OpenAI's cart pole problem.

Notifications You must be signed in to change notification settings

sidlakkoju/Cartpole-DQN

Repository files navigation

CartPole_DeepLearning

Based on Deepmind's paper "Playing Atari with Deep Reinforcement Learning"

Sequential neural network power by TensorFlow and Keras.


View the trained DQN in action:
Trained Cartpole DQN in action:

Note, cartpole_training_2 contains the weights which allow the model to perform the best. These weights were generated after 150 training episodes with no exploration. Feel free to experiment with the model and the training parameters! :)

Releases

No releases published

Packages

No packages published

Languages