/tf-meetup-reinforce

Slides and code for the introduction to Reinforcement Learning session at the Barcelona TensorFlow Meetup (October 4, 2018)

MIT LicenseMIT

Introduction to Policy Gradients with TensorFlow (Barcelona TensorFlow Meetup)

Resources for the introduction to Reinforcement Learning session at the Barcelona TensorFlow Meetup (October 4, 2018). They cover a quick introduction to RL, REINFORCE and code example to solve the CartPole task using TensorFlow Eager.

Resources

Results

plot

After a while, the agent is able to keep the pole in balance during the whole episode (termination is forced after 200 steps):

visualization