Skip to content

Repository for work related to Udacity's Deep Reinforcment Learning Nanodegree.

Notifications You must be signed in to change notification settings

hoseasiu/DeepRLNanodegree

Repository files navigation

DeepRLNanodegree

Repository for work related to Udacity's Deep Reinforcment Learning Nanodegree. See individual project subfolders for more detailed descriptions.

NavigationProject

Train an agent to gather bananas in the Unity Banana Collector environment. Demonstrates the Deep Q-Learning algorithm and the effects of various hyperparameter settings. Note that this version of the agent does not use image information from the environment, but rather uses state information directly in a 37-dimensional vector that includes ego-information and ray-based information about its surroundings.

ContinuousControl

Train a robotic arm to follow a moving target region in space. Demonstrates the Deep Deterministic Policy Gradient Algorithm - an actor-critic method for continuous action spaces.

Multi-Agent Reinforcement Learning

Train two agents to play a game of tennis in the Unity Tennis environment. Demonstrates DDPG with prioritized experience replay with multiple agents.

About

Repository for work related to Udacity's Deep Reinforcment Learning Nanodegree.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published