Jupyter Notebook art enhaning pipe with scripts to train Stylegan2 models on new data from scratch or via transfer learning. Also contains scripts for generating images from trained models, and projecting images onto the generatable manifold.This notebook is an extension of Mikael Christensen's notebook, and includes his scripts for image generation and projection.
Theoretical part:
- Tasks and methods of optimization
- Lagrange multiplier method
- Newton's method
Video: https://youtu.be/TXPdZTJ36fY