/drlnd_continuous-control

Train an agent to solve the Reacher Unity ML-Agents Environment

Primary LanguagePythonBSD 3-Clause "New" or "Revised" LicenseBSD-3-Clause

Solving the DRLND Unity Reacher Environment

Train an agent to solve the Reacher Unity Environment from the Deep Reinforcement Learning Nanodegree on Udacity.

  • TODO: Train for ~500000 episodes and save model
    • set configs
      • config.max_steps = 500000
      • config.save_interval = int(1e5)
  • TODO: Add trained example run
  • TODO: Add reward plot
  • TODO: Add description and link detailed report

Prerequisites

  • conda or miniconda (recommended)
  • make
  • Download the environment that matches your OS following the Getting Started from the DRLND repo and unpack it in the root of this project

Install Environment

Automated Install

Simply run make install to install all requirements in a conda environment called drlnd_control.

Manual Install

Create a conda environment called drlnd_control with Python3.6 and activate it using the following commands

conda create --name drlnd_control python=3.6
conda activate drlnd_control

Then install the requirements file requirements.txt and install the drlnd_control ipykernel.

pip install -r $(PWD)/requirements.txt
python -m ipykernel install --user --name drlnd_control --display-name "drlnd_control"

Run the Code

Next, run make start to start the Jupyter notebook server and use your favorite browser to navigate to http://localhost:8888/?token=abcd.

Train an Agent

  • TODO: Docs how to train an agent

Watch a Trained Agent

  • TODO: Docs how to watch a successful agent