/cartpole-dqn

OpenAI Gym cartpole solved by a Neural Network (DQN) in Tensorflow 2

Primary LanguagePythonMIT LicenseMIT

OpenAI Gym cartpole solved by a Neural Network (DQN) in Tensorflow 2

CartPole-v0 defines "solving" as getting average reward of 195.0 over 100 consecutive trials. Run python play.py to check.

Requirements

  • Python 3.6
  • Pipenv

Installation

$ pipenv install
$ pipenv shell

Create/Train/Play with Model

$ pipenv run python train.py
$ pipenv run python play.py

Show log

$ tensorboard --logdir=logs

Currently I get an error: tensorflow/tensorflow#32384

Resources