This repository provides a deep learning framework with implementations of classic neural network architectures. It includes Restricted Boltzmann Machines with continuous-valued variants, Deep Belief Networks, denoising autoencoders and their stacked variants, logistic regression, hidden layers, multilayer perceptrons with dropout, and convolutional neural networks. The library is written primarily in Java with support for multiple languages, and is based on foundational deep learning research from Bengio, Larochelle, and Vincent.