I liked GANs for a while. So much, that I started writing about them in reputable publications before I even had my first job in the Machine Learning field.
I trained them mostly on faces, trying to enter the Uncanny Valley, freaking some people out in the process. Then I lost interest – forgot about them entirely.
I followed GAN artists for a while, people working at the verge of technology and art, and it fascinates me. So when I found out how to download more than 15’000 images from a clothing retailer, I took the chance and created a dataset that I could use for my new project.
I scraped 15K+ pieces of brand clothing off a major clothing store (psshh, don’t tell them) to train a GAN and see if it can generate some cool designs.
🤖 So far, human designers have nothing to worry about.
— Dominic Monn (@dqmonn) November 12, 2018
I pushed this update to my personal Twitter at the end of last year – and people loved to see it. As I learn more about GANs, I’d like to share it in this blog, together with my discoveries of the latent space and finally the great products I can create, based on logic, coupled with art and imagination.