1、引言
主要创新:
1)complementary search techniques
2)new efficient versions of nonlinearities practical for mobile setting
3)new efficient network design
4)a new efficient segmentation decoder
2、相关工作
1)novel handcrafted structures
SqueezeNet、MobileNet、ShuffleNet、CondenseNet、ShiftNet
2)algorithmic neural architecture search
3)quantization
4)knowledge distillation
3、高效的mobile building blocks
1)MobileNet-V1:depth-wise separable convolution
2)MobileNet-V2:linear bottleneck and inverted residual
3)MnasNet:light attention modules based on squeeze and excitation into the bottleneck
4、网络搜索
1)全局网络结构(block-wise):platform-aware NAS(network architecture search)
2)每层filters的数量(layer-wise):NetAdapt algorithm
5、网络提升
主要创新:重新设计网络最开始和最后computionally-expensive 的层,介绍了一种新的nonlinearity-h-swish,计算更快且对于量化更加友好。
1)重新设计expensive layers
之前的模型一般都将1*1 conv作为最终的layer,以展开到一个更高维的特征空间
原文:https://www.cnblogs.com/wt-seu/p/12401257.html