Learn to be a painter using Neural Style Painting
summary
Uses VGG-19 to extract content and style features and optimizes a loss made of content loss style loss and total variation denoising, tuning the alpha/beta content vs style weights by launching many independent runs. Implementation uses TensorFlow + SparkMagic + Livy, spawning subprocess jobs on executors to parallelize hyperparameter sweeps. Example: a 12‑configuration run on a single master/worker r4 cluster takes about nine hours and produces stylized images (Space Needle examples) without GPU acceleration.
AI-written summary · about
chapters (10)
- 0:00Neural painting intro
- 1:27Analytical workflow recap
- 12:50Convolutional neural networks
- 17:03TensorFlow data flow
- 18:28Livy Spark Magic
- 22:24Neural Style Transfer explained
- 27:26Style transfer demo
- 30:16Scaling with Spark TensorFlow
- 37:47Results and applications
- 38:53Performance and future
related talks
25mMixing art with Python- an introduction to Style Transfer
58mStyle Transfer AI - How Neural Networks Generate Art
13m[Interactive] Neural Style Transfer: The Art and Science of Generative AI
37mLearning AI with Van Gogh, Botero and Cruz‑Diez
29mIntegrating Stable Diffusion Generative AI in Artist Workflows
50mVan Gogh Meets Neural Networks
15mProgramming Visual Art
57mWhat if Picasso was a robot? Teaching art to the machine