Gen planning

This is a planning doc for generist art creation. For detailed how-tos and writeup, see other pubs in this group.

Toolbox of neural networks for image synthesis

• https://github.com/wxs/subjective-functions

• https://github.com/AaronLeong/BigGAN-pytorch

• https://github.com/lengstrom/fast-style-transfer

• https://github.com/google/deepdream/blob/master/dream.ipynb

• https://tfhub.dev/deepmind/biggan-512/2

• picbreeder, ganbreeder?

Customize BigGAN

1. Take pretrained bigGAN generator

2. Add an extra layer

3. See layer with Met objects / objectives

—-> an image synthesis sandbox !

Specialize training data

• use Met images

• use wikiart genres? scraper:
  https://github.com/robbiebarrat/art-DCGAN/blob/master/genre-scraper.py

May not need to if it suffices to specialize bigGAN with Met images
But could just choose specific genres {or category combinations}

Biggan categories == imagenet categories == https://gist.github.com/yrevar/942d3a0ac09ec9e5eb3a