SparseMask

Pytorch implementation of "Exploring Fine-Grained Sparsity in Convolutional Neural Networks for Efficient Inference", TPAMI 2022

[Paper]

Motivation

• Connection between the sparsity in human brains and the sparsity in CNNs

• Feature sparsity in CNNs

Overview

• Sparse Mask Generation

• Sparse Mask Convolution

Applications

1. Point Cloud Semgantic Segmentation [code]

• Network Architecture

• Results

2. Singe Image Super-Resolution [code]

• Network Architecture

• Results

3. Stereo Matching

• Network Architecture

• Results

Citation

@Article{Wang2022Exploring,
  author  = {Longguang Wang and Yulan Guo and Xiaoyu Dong and Yingqian Wang and Xinyi Ying and Zaiping Lin and Wei An},
  title   = {Exploring Fine-Grained Sparsity in Convolutional Neural Networks for Efficient Inference},
  journal = {IEEE Transactions on Pattern Analysis and Machine Intelligence},
  year    = {2022},
}