PyTorch implementation of Dynamic Grouping Convolution and Groupable ConvNet in Differentiable Learning-to-Group Channels via Groupable Convolutional Neural Networks.
- Kronecker Product is utilized to construct the sparse matrix efficiently and regularly.
- Discrete optimization is solved with the Straight-Through Estimator trick.
- Automatically learn the number of groups in an end-to-end differentiable fashion.
DGConv is used as a drop-in replacement of depthwise separable convolution in the original ResNeXt to build G-ResNeXt-50/101 network architectures. Here are some results of their performance comparison.
| Architecture | LR decay strategy | Top-1 / Top-5 Accuracy |
|---|---|---|
| ResNeXt-50 (32x4d) | cosine (120 epochs) | 78.198 / 93.916 |
| G-ResNeXt | cosine (120 epochs) | 78.592 / 94.106 |
@InProceedings{Zhang_2019_ICCV,
author = {Zhang, Zhaoyang and Li, Jingyu and Shao, Wenqi and Peng, Zhanglin and Zhang, Ruimao and Wang, Xiaogang and Luo, Ping},
title = {Differentiable Learning-to-Group Channels via Groupable Convolutional Neural Networks},
booktitle = {The IEEE International Conference on Computer Vision (ICCV)},
month = {Oct},
year = {2019}
}