Putting 3D Spatially Sparse Networks on a Diet

TL;DR

This paper investigates weight sparsity in 3D neural networks and introduces a compact, sparse 3D convolutional network that maintains performance while significantly reducing size and computational cost.

Contribution

It is the first comprehensive study on weight sparsity in spatially sparse 3D networks and proposes a novel WS^3-Convnet with high compression and efficiency.

Findings

Achieves 99% parameter reduction with only 2.15% performance drop

Reduces computational cost by 95%

Speeds up inference by 45%

Abstract

3D neural networks have become prevalent for many 3D vision tasks including object detection, segmentation, registration, and various perception tasks for 3D inputs. However, due to the sparsity and irregularity of 3D data, custom 3D operators or network designs have been the primary focus of research, while the size of networks or efficacy of parameters has been overlooked. In this work, we perform the first comprehensive study on the weight sparsity of spatially sparse 3D convolutional networks and propose a compact weight-sparse and spatially sparse 3D convnet (WS^3-Convnet) for semantic and instance segmentation on the real-world indoor and outdoor datasets. We employ various network pruning strategies to find compact networks and show our WS^3-Convnet achieves minimal loss in performance (2.15\% drop) with orders-of-magnitude smaller number of parameters (99\% compression rate) and computational cost (95\% reduction). Finally, we systematically analyze the compression patterns of WS^3-Convnet and show interesting emerging sparsity patterns common in our compressed networks to further speed up inference (45\% faster). \keywords{Efficient network architecture, Network pruning, 3D scene segmentation, Spatially sparse convolution}