In Spring 2020, as NYC was on a COVID-19 lockdown, I took an online class organized by the wonderful folks at SSHH taught by Derrick Schultz on how to use RunwayML to create AI art using GANs.

This included learning how to train a Generative Adversarial Network (GAN), style transfer, creating datasets, chaining models, p5.js interpolations, and much more.

Incidentally, this was before the explosion of generative AI and LLMs in ~2022. So it is very possible / probable that a lot of what I learned is completely out of date now. 🤷🏽‍♂️

That being said, this was still a very fun time!

Here's a few images and videos I created:

img0064, MicroGAN, trained on Nikon Photomicrography Competition submissions

img0279, MicroGAN, trained on Nikon Photomicrography Competition submissions

img0446, MicroGAN, trained on Nikon Photomicrography Competition submissions

img0485, MicroGAN, trained on Nikon Photomicrography Competition submissions

img0595, MicroGAN, trained on Nikon Photomicrography Competition submissions

img0623, MicroGAN, trained on Nikon Photomicrography Competition submissions

img0734, MicroGAN, trained on Nikon Photomicrography Competition submissions

img0744, MicroGAN, trained on Nikon Photomicrography Competition submissions

img0779, MicroGAN, trained on Nikon Photomicrography Competition submissions

img0953, MicroGAN, trained on Nikon Photomicrography Competition submissions

img0979, MicroGAN, trained on Nikon Photomicrography Competition submissions

img0995, MicroGAN, trained on Nikon Photomicrography Competition submissions