Lossy Compression using Deep Neural Networks Some models / experiments for compression on MNIST, CIFAR and Text data. Implementation of birary quantized representation models.