Gearloose

The Gearloose is a multi-disciplinary project that builds a workflow for Generative Adversarial Network (GAN). Gearloose consits of tools preparing a dataSet from fetching, normalizing, to training a customed machine learning model. Built models could later generate new dataSets independetly.

Workflow:

1. Fetching a DataSet
2. Normalizing a DataSet
3. Training a custom model of GAN
4. Generating data using a trained GAN model

DataSet Tools

This is a package of tools for quickly normalizing image datasets to be ready for a Generative Adversarial Network. See its own documentation here

Training a custom model of Generative Adversarial Network.

Generating data from already trained model