Official code for Hierarchical Attribute CNNs (hCNNs). hCNNs are highly structured CNNs that formulate each layer as a multi-dimensional convolution. hCNNs provide a framework that allows to study and understand mathematical and semantic properties of deep convolutional networks.

Reference: J.-H. Jacobsen, E. Oyallon, S. Mallat, A.W.M. Smeulders; Hierarchical Attribute CNNs. Proceedings of the ICML 2017 Principled Approaches to Deep Learning Workshop

Tested with: cudnn v5.1; cuda 8.0; Tensorflow 0.12; Keras 1.2.2

1. Understanding Deep Convolutional Networks, Mallat, 2016