VAE-Tensorflow

Implemented variational auto-encoder with MNIST dataset using TensorFlow.

Prerequisites

• Tensorflow
• Scipy

Usage

python main.py Training process is activated, as parameters would be saved as checkpoint file. After every 10 epochs, generated MNIST number will be saved as png file.

References

• y0ast/VAE-TensorFlow
  • I got intuition of variational auto-encoder implementation in this repository.
• cdoersch/vae-tutorial
  • This repository is supplemented code to the Tutorial on Variational Autoencoders. This code is Caffe version of variational auto-encoder for MNIST dataset.