Implementation of recent Deep Learning papers

Papers

• Densely Connected Convolutional Network implemented in the DenseNet folder

• Visualizing and Understanding Convolutional Networks implemented in the DeconvNet folder

• Improving Stochastic Gradient Descent With Feedback implemented in the Eve folder

• Colorful Image Colorization implemented in the Colorful folder

• Deep Feature Interpolation for Image Content Changes implemented in the DFI folder

• Several Generative Adversarial Networks (GAN) models and techniques in the GAN folder

• pix2pix in the pix2pix folder

• InfoGAN in the InfoGAN folder

• WassersteinGAN in the WassersteinGAN folder and WGAN-GP folder for a tensorflow implementation.

• BEGAN in the BEGAN folder

• Scaling the Scattering Transform: Deep Hybrid Networks in the ScatteringTransform folder

• Sobolev Training for Neural Networks in the Sobolev folder

• Self-Normalizing Networks in the SELU folder