Deep Convolutional Generative Adversarial Networks (DCGAN)

INFO

I’ve worked with several datasets and used Deep Convolutional Generative Adversarial Networks (DCGAN) for data generation. While the model was trained on many different configurations, the following results are from a basic example with just 50 epochs. These results and graphs are a sample of the model’s output, demonstrating how the network progressively learns and generates realistic data. Among my many trainings, this is a simpler example to highlight the potential of the technique.

Training

Process

Outcomes

Graph of the DCGAN loss vs. epoch