This is the code for "How to Generate Images - Intro to Deep Learning #14' by Siraj Raval on YouTube
This weeks challenge is to use a VAE to generate images of something other than MNIST! You can use Keras for your code as well. Bonus points if you generate video like this.
This is the code for this video on Youtube by Siraj Raval as part of the Udacity Deep Learning Nanodegree. We're using a variational autoencoder to generate novel digit images after training on the MNIST dataset. We use Keras in this repository so the code is relatively simple, for a more in depth look (at say, reparameterization check this repo).
- numpy
- keras
- scipy
- matplotlib
Install dependencies using pip.
Run jupyter notebook in terminal and the code will pop up in your browser. Install it here if you don't have it.
Credits go to the creator of Keras. I've merely created a wrapper to get people started.