This repo is a solution to Udacity Deep Reinforcement Learning Nano Degree project Colab and Compete

Interestingly the agent has discovered the dexterity of a double shot in Tennis.

• Environment: Two agents control rackets to bounce a ball over a net.
• Goal: Each agent should keep the ball in play.
• Reward:
  • +0.1: When an agent hits the ball over the net.
  • -0.01: When an agent lets a ball hit the ground or hits the ball out of bounds.
• State space: 8*3 variables corresponding to position and velocity of the ball and racket. Each agent receives its own, local observation.
• Action space: 2 continuous values, corresponding to movement toward (or away from) the net, and jumping.
• Solved: When agents gets an average score of +0.5 (over 100 consecutive episodes, after taking the maximum over both agents).

brew install python3 swig && \
    brew install opencv3 --with-python && \
    pip3 install --upgrade pip setuptools wheel

sudo apt-get install swig python3 python3-venv

python3 -m venv .venv && \
    source .venv/bin/activate && \
    pip install -r requirements.txt

1. Download "Tennis" environment based on your machine.
   • Linux: click here
   • Linux Headless: click here
   • Mac OSX: click here
   • Windows (32-bit): click here
   • Windows (64-bit): click here
2. Place the file in the env directory, and unzip the file.

source .venv/bin/activate

python3 train.py

Required checkpoints are already available in checkpoints/ directory.

python3 test.py