Bored monkeys have become a symbol of NFT mania. Terribly ugly and ridiculously expensive, their value stems from an artificial shortage.
It would be a great shame if someone offered endless free versions of collectibles – exactly what Yannick Kilcher did.
The machine learning expert has created an application that generates as many bored monkeys as your heart desires.
The system is powered by generative competition networks (GANs), which create new content by pitting two neural networks against each other: a generator that produces new images and a discriminator that guesses which ones are fake.
Over time, the generator begins to create new photos that can pass as authentic.
Kilcher first scraped the Bored Ape collection of 10,000 images.
He then trained Nvidia StyleGAN2 neural network in the data set.
The finished system produces Bored Apes, which are completely unique – and interchangeable.
However, Kilcher admits that the images are not always perfect.
“The quality of my system is not very good, there are often visible artifacts,” he told TNW.
Still, Kilcher has created something really original: an app that generates monkeys based on images you upload.
The application uses OpenAI’s CLIP – the same model that powers the DALL-E 2 image generator. You can try it yourself, but you’ll need a local graphics processor to boot the system.
Our thoughts and prayers are focused on Jimmy Fallon and Paris Hilton at this difficult time.