GenSynth Documentation

Before you Begin

This document explains how to prepare your datasets and models for use with GenSynth, and how to configure a software interface to your datasets, known as the Dataset Interface plugin.

It assumes you have already installed GenSynth and gained a basic familiarity with the terminology by working through the tutorial.

Before starting your data preparation, gather information about the model and training optimizer.

If you do not have these items prepared, refer to the GenSynth Installation Guide.

To complete a GenSynth project to optimize or explain a deep-learning model, you will need to be prepared to:

  • Make data available for GenSynth to use in training and validation.

  • Identify the input, output, loss, and metric tensors of your model.

  • Provide an algorithm for measuring the performance quality of model variants created by GenSynth.

Once you have configured each of these aspects, you will be able to run multiple iterations of GenSynth to gain insight about your model and to improve it.