Use containers to create software environments that you can use exactly where you need them, whether that’s desktop, server, or cloud environments. Because you install only the software libraries and packages you need for your applications, containers are lightweight and provide a reproducible and reliable way of sharing applications without worrying about configuring installations each time. You can dockerize MATLAB to integrate it with a Continuous Integration/Continuous Delivery (CI/CD) pipeline that is completely containerized. MATLAB has integrations for CircleCI^®, Jenkins^®, and Travis CI.

The MATLAB Deep Learning Container is designed to take full advantage of high-performance NVIDIA^® GPUs to speed up your deep learning applications. You can access the MATLAB Deep Learning Container from anywhere using a web browser or VNC connection.

You can create your own container image with a customized MATLAB installation using the reference Dockerfile on GitHub^®. The README file in the GitHub repository shows you how to install MATLAB toolboxes in the container image.

You can also package MATLAB applications in a Docker^® container. The resulting image has a much smaller size than that of a fully containerized MATLAB.