This is caffe implementation of ShuffleNet, For details, please read the original paper:
"ShuffleNet: An Extremely Efficient Convolutional
Neural Network for Mobile Devices" by Xiangyu Zhang et. al. 2017.
This code is based on camel007's implementation(https://github.com/camel007/Caffe-ShuffleNet), but I recode the cuda file for acceleration.
message LayerParameter {
...
optional ShuffleChannelParameter shuffle_channel_param = 164;
...
}
...
message ShuffleChannelParameter {
optional uint32 group = 1[default = 1]; // The number of group
}
The top-1/5 accuracy rates by using single center crop (crop size: 224x224, image size: 256x256):
Network | Top-1 | Top-5 | Download | Architecture |
---|---|---|---|---|
ShuffleNet | 62.8% | 84.7% | caffemodel (7.0 MB) | deploy |