SparseNet

by Lyken17

Lyken17 / SparseNet

[ECCV 2018] Sparsely Aggreagated Convolutional Networks https://arxiv.org/abs/1801.05895

127 Stars 27 Forks Last release: Not found MIT License 34 Commits 0 Releases

Available items

No Items, yet!

The developer of this repository has not created any items for sale yet. Need a bug fixed? Help with integration? A different license? Create a request here:

SparseNet

Sparsely Aggregated Convolutional Networks [PDF]

Ligeng Zhu, Ruizhi Deng, Michael Maire, Zhiwei Deng, Greg Mori, Ping Tan

What is SparseNet?

SparseNet is a network architecture that only aggregates previous layers with exponential offset, for example, i - 1, i - 2, i - 4, i - 8, i - 16 ...

Why use SparseNet?

The connectivity pattern yields state-of-the-art arruacies on small dataset CIFAR/10/100. On large scale ILSVRC 2012 (ImageNet) dataset, SparseNet achieves similar accuracy as ResNet and DenseNet, while only using much less parameters.

Better Performance

Without BC With BC
Architecture Params CIFAR 100
DenseNet-40-12 1.1M 24.79
DenseNet-100-12 7.2M 20.97
DenseNet-100-24 28.28M 19.61
--- --- ---
SparseNet-40-24 0.76M 24.65
SparseNet-100-36 5.65M 20.50
SparseNet-100-{16,32,64} 7.22M 19.49
Architecture Params CIFAR 100
DenseNet-100-12 0.8M 22.62
DenseNet-250-24 15.3M 17,6
DenseNet-190-40 25.6M 17.53
--- --- ---
SparseNet-100-24 1.46M 22.12
SparseNet-100-{16,32,64} 4.38M 19.71
SparseNet-100-{32,64,128} 16.72M 17.71

Efficient Parameter Utilization

  • Parameter efficiency on CIFAR

  • Paramter efficiency on ImageNet

We notice sparsenet shows comparable efficiency even compared with pruned models.

Pretrained model

Refer for source folder.

Cite

If SparseNet helps your research, please cite our work :)

@article{DBLP:journals/corr/abs-1801-05895,
  author    = {Ligeng Zhu and
               Ruizhi Deng and
               Michael Maire and
               Zhiwei Deng and
               Greg Mori and
               Ping Tan},
  title     = {Sparsely Aggregated Convolutional Networks},
  journal   = {CoRR},
  volume    = {abs/1801.05895},
  year      = {2018},
  url       = {http://arxiv.org/abs/1801.05895},
  archivePrefix = {arXiv},
  eprint    = {1801.05895},
  biburl    = {https://dblp.org/rec/bib/journals/corr/abs-1801-05895},
  bibsource = {dblp computer science bibliography, https://dblp.org}
}

We use cookies. If you continue to browse the site, you agree to the use of cookies. For more information on our use of cookies please see our Privacy Policy.