View Code? Open in Web Editor NEW

An environment for interactive exploration of reinforcement learning

License: BSD 3-Clause "New" or "Revised" License

Python 45.78% Jupyter Notebook 6.80% RobotFramework 18.36% Shell 10.08% Batchfile 2.28% Jinja 16.69%

gt-coar-lab's Introduction

GTCOARLab

an environment for interactive exploration of reinforcement learning

Build	Interactive Demo

Downloading GTCOARLab

View the releases on GitHub
Choose the right artifacts for your operating system and hardware
- CPU: does not require a CUDA-capable graphics card and drivers
- GPU: some features require a CUDA-capable graphics card and drivers
Download the artifacts (they are big, it will take a while)
- If your platform contains multiple .zip and .z01, .z02 files, you need to download them all
  - a GUI or the zip command line tool will know how to decompress them correctly by using the .zip file, which should be smaller than the others
Follow the instructions for e.g. miniconda
- on Windows, prefer installing to a short path on a fast device (e.g. SSD) such as C:/gtcl
- if you've already installed Anaconda, Miniconda, or Miniforge, it is recommended to not enable shell integration, environment variables, or registering as the default Python
From the command line, activate the environment
- on Windows, you should see a Start Menu entry
- run jupyter lab, and you should see the JupyterLab interface open in your default browser with
  - to choose your browser, start jupyter lab --no-browser and copy/paste the URL shown into your browser of choice
If you run into problems, create an issue on GitHub