Hierarchical Block Sparse Neural Networks

Abstract: Sparse deep neural networks(DNNs) are efficient in both memory and compute when compared to dense DNNs. But due to irregularity in computation of sparse DNNs, their efficiencies are much lower than that of dense DNNs on regular parallel hardware such as TPU. This inefficiency leads to poor/no performance benefits for sparse DNNs. Performance issue for sparse DNNs can be alleviated by bringing structure to the sparsity and leveraging it for improving runtime efficiency. But such structural constraints often lead to suboptimal accuracies. In this work, we jointly address both accuracy and performance of sparse DNNs using our proposed class of sparse neural networks called HBsNN (Hierarchical Block sparse Neural Networks). For a given sparsity, HBsNN models achieve better runtime performance than unstructured sparse models and better accuracy than highly structured sparse models.