Skip to content

Code to train Reacher environment using the Unity ML-Agents as a part for the second project of the Deep Reinforcement Learning Nanodegree.

Notifications You must be signed in to change notification settings

Tomas0413/ContinuousControl

Repository files navigation

Project Details

The agent is a double-jointed arm. The goal is to move to a target location.

  • States: 33
  • Actions: 4 (torque applicable to two joints)
  • Rewards: +0.04 for each step that the agent's hand moves toward the goal location

Getting Started

Open ContinuousControl.ipynb Jupyter notebook, it also contains a section on how to install the necessary dependencies.

Download the Reacher environment (binary file) for your operating system and configure file_name variable (in ContinuousControl.ipynb Jupyter notebook) to point to the location of the binary file.

Instructions

Open ContinuousControl.ipynb Jupyter notebook and execute each code block to train the agent using DDPG.

About

Code to train Reacher environment using the Unity ML-Agents as a part for the second project of the Deep Reinforcement Learning Nanodegree.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published