The deep learning model has many parameters (the model is very large), which is a big bottleneck restricting the deployment of deep learning models on mobile terminals. Recently, many lightweight deep learning models have been proposed. The following are some classic lightweight deep learning models. Summarize:
1、Squeezenet:Alexnet-level accuracy with 50x fewer parameters and¡ 1mb model size.
arXiv preprint arXiv:1602.07360, 2016
2、MobileNets:Efficient Convolutional Neural Networks for Mobile Vision Applications
arXiv:1704.04861v1 [cs.CV] 17 Apr 2017
3、Densely ConnectedConvolutional Networks
arXiv:1608.06993v4 [cs.CV] 27 Aug 2017
4、Xnornet: Imagenetclassification using binary convolutional neural networks.
arXiv preprint arXiv:1603.05279, 2016
5、Quantized convolutionalneural networks for mobile devices
arXiv preprint arXiv:1512.06473, 2015
6、Xception: DeepLearning with Depthwise Separable Convolutions
arXiv:1610.02357v3 [cs.CV] 4 Apr 2017
7、ProjectionNet: LearningEfficient On-Device Deep Networks Using Neural Projections
arXiv:1708.00630v2 [cs.LG] 9 Aug 2017
8、Factorizedconvolutional neural networks
arXiv preprint arXiv:1608.04337, 2016
9、LearningTransferable Architectures for Scalable Image Recognition
atXiv:1707.07012v2 [cs.CV] 25 Oct 2017
10、Structured transformsfor small-footprint deep learning.
11、Deep compression:Compressing deep neural network with pruning, trained quantization and huffmancoding
12、Quantized neuralnetworks: Training neural networks with low precision weights and activations.
arXivpreprint arXiv: 1609.07061, 2016