Rework of Deeplearning4J Examples

In order to make the example repository more useful, we have reworked and restructured it entirely.

The examples are now split into several separate projects. The previous multi module approach made it hard for many interested users to understand the build process and lead to unnecessary confusion.

With this change, you can directly import the project that you are interested in, and don’t have to worry about multi-module maven builds. And it also reduces the size of the required dependency downloads. You only have to download what you’ll actually use.

The projects are based on the functionality of the included examples now, instead of being grouped by library. This makes it easier to follow through the examples, as usually many of the libraries are used together in real world applications.

As some examples follow more of a quickstart theme, while others show how to use advanced features, we have also split the examples within a project along those lines as well. This should also give you an intuition for what is considered to be something that you should learn first (“quickstart”), and what is considered to be something that you should only need once you graduate to more complex problems (“advanced”).

Each project now also comes with a detailed README file that for each example explains what it showcases.

If you always wanted to learn Deeplearning4J but hesitated because it seemed overwhelming, it may be a good time to try it out now.

Go to the examples repository. or head over to the Examples Tour in the documentation.

If you run into issues, feel free to ask questions here on the community forums.