Tutorial

The goal of this tutorial is to show how create a JAW project, i.e. a project with the JAW template architecture and some built-in methods. You can also decide to keep the architecture but without using the JAW package. The most important is the architecture.

For made this example more concrete, we will reimplement this tutorial in our own way (credits to Jeremy Fix). If you want know more about the machine learning, we strongly recommend this webbook, especially this definition. We can find the final example project inside the folder src/example/.

Note

This tutorial hasn’t the pretention to teach you how write a better network model, but only to provide a base of code for neophytes or peoples looking for a base for their own architectures.

• Getting started
  • Project generation
  • File tree
• Data preprocessing
  • Load dataset
  • Dataset to data loaders
• Training functions
  • Train
  • Eval
• Model
  • First model
  • Regularization
• Custom loss
• Training Process
  • Write the process
  • Define launch options
  • Launch the training