Generated images are under images.
Generated names are under names.
The models can be found here
The code to make the template can be found here
All the preprocessed images can be found here
It produces images which look exactly like how a Pokémon would look, but the Pokémon's do not actually exist. The model has successfully learn a latent space of different Pokémon like species by training on a corpus of "real" Pokémon images. We then combine these processed images into the card layout to make a new card based on the Pokémon. We also trained a minGPT on existing Pokémon names and asked it to generate new names! Magale, Pimate, Garenige and Popet are some of the Pokémon names that don't exist.
We built it with the help of variation of StyleGAN model which we fine-tuned to generate images. We trained a minGPT model(a minimal implemetation of GPT by Karpathy). We then manually prepared the cards(of course we used python, duh) and put them on a website.