This project contains an agent based on Deep Reinforcement Learning that can learn from zero (any labeled data) to collect yellow bananas instead of blue bananas in a vast, square world.
It's use the Unity Machine Learning Agents Toolkit to design, train, and evaluate deep reinforcement learning algorithms implementations.
The environment used for this project is the Udacity version of the Banana Collector environment, from Unity. The goal of the agent is to collect as many yellow bananas as possible while avoiding blue bananas.
The installation process is divided in three parts:
- Python 3.6.x
- Dependencies
- Unity's Environment
To install and run this project you will need to have Python 3.6.x.
You could download olds Anaconda versions here
The reason for it is that Unity ml-agents needs TensorFlow in version 1.15 and TensorFlow 1.15 needs Python. 3.6.x.
All packages and their versions are describe in requirements.txt.
You need to run these commands before to install all dependencies:
pip install --upgrade pip
pip install -r requirements.txt
pip -q install ./python
You'll have these three commands in the first part of [Navigation.ipynb], as well.
You need only select the environment that matches your operating system:
Linux: click here Mac OSX: click here Windows (32-bit): click here Windows (64-bit): click here
(For AWS) If you'd like to train the agent on AWS (and have not enabled a virtual screen), then please use this link to obtain the environment.
Then, place the unzipped folder on the root of this repository.
The contents of this repository are covered under the MIT License.