The Joy of Neural Painting

TL;DR

This paper introduces a transfer learning approach to significantly accelerate the training of neural painting models based on GANs, achieving high-quality artistic outputs in hours instead of days.

Contribution

It presents a novel transfer learning method that reduces training time for neural painters without compromising visual quality.

Findings

Training time reduced from days to hours

Achieved comparable visual aesthetics

Effective transfer learning application

Abstract

Neural Painters is a class of models that follows a GAN framework to generate brushstrokes, which are then composed to create paintings. GANs are great generative models for AI Art but they are known to be notoriously difficult to train. To overcome GAN's limitations and to speed up the Neural Painter training, we applied Transfer Learning to the process reducing it from days to only hours, while achieving the same level of visual aesthetics in the final paintings generated. We report our approach and results in this work.