Machine Learning

I began playing with some of the online Deep Dream sites. They are sites that let you upload a photo or image and "apply" a style to it based upon a trained model. So, you can turn your image into a "Monet." I decided I preferred to use all my own work. How to accomplish that using Tensorflow is very difficult to understand. Yet, I found a simple way using tools/code available online. I used what I think is a watered down method for style transfer using a script(include?) tag for Javascript. So, I used my own source image, one of my pieces for the style to apply, and the image to the right is the AI result. It should be noted that I think that's the process I used. Thanks Google! Here's the link: https://github.com/tensorflow/magenta-js/tree/master/image#example-applications

To apply a style to the left three photographs, I used Google's Colab and CySmith's Neural Style Transfer. I switched to that from Javascript since it offers many more options. For instance you can increase the number of iterations for learning. And you can use many images for that learning. You can vary the file size.