ShuffleNet V2: Practical Guidelines for Efficient CNN Architecture Design
论文链接：https://arxiv.org/pdf/1807.11164.pdf
PyTorch：https://github.com/shanglianlm0525/PyTorch-Networks
ShuffleNet V2 unit
ShuffleNet v2 architecture
PyTorch代码
import torch
import torch.nn as nn
import torchvision
def Conv3x3BNReLU(in_channels,out_channels,stride,groups):
return nn.Sequential(
nn.Conv2d(in_channels=in_channels, out_channels=out_channels, kernel_size=3, stride=stride, padding=1,groups=groups),
nn.BatchNorm2d(out_channels),
nn.ReLU6(inplace=True)
)
def Conv3x3BN(in_channels,out_channels,stride,groups):
return nn.Sequential(
nn.Conv2d(in_channels=in_channels, out_channels=out_channels, kernel_size=3, stride=stride, padding=1,groups=groups),
nn.BatchNorm2d(out_channels)
)
def Conv1x1BNReLU(in_channels,out_channels):
return nn.Sequential(
nn.Conv2d(in_channels=in_channels, out_channels=out_channels, kernel_size=1, stride=1),
nn.BatchNorm2d(out_channels),
nn.ReLU6(inplace=True)
)
def Conv1x1BN(in_channels,out_channels):
return nn.Sequential(
nn.Conv2d(in_channels=in_channels, out_channels=out_channels, kernel_size=1, stride=1),
nn.BatchNorm2d(out_channels)
)
class HalfSplit(nn.Module):
def __init__(self, dim=0, first_half=True):
super(HalfSplit, self).__init__()
self.first_half = first_half
self.dim = dim
def forward(self, input):
splits = torch.chunk(input, 2, dim=self.dim)
return splits[0] if self.first_half else splits[1]
class ChannelShuffle(nn.Module):
def __init__(self, groups):
super(ChannelShuffle, self).__init__()
self.groups = groups
def forward(self, x):
'''Channel shuffle: [N,C,H,W] -> [N,g,C/g,H,W] -> [N,C/g,g,H,w] -> [N,C,H,W]'''
N, C, H, W = x.size()
g = self.groups
return x.view(N, g, int(C / g), H, W).permute(0, 2, 1, 3, 4).contiguous().view(N, C, H, W)
class ShuffleNetUnits(nn.Module):
def __init__(self, in_channels, out_channels, stride, groups):
super(ShuffleNetUnits, self).__init__()
self.stride = stride
if self.stride > 1:
mid_channels = out_channels - in_channels
else:
mid_channels = out_channels // 2
in_channels = mid_channels
self.first_half = HalfSplit(dim=1, first_half=True)
self.second_split = HalfSplit(dim=1, first_half=False)
self.bottleneck = nn.Sequential(
Conv1x1BNReLU(in_channels, in_channels),
Conv3x3BN(in_channels, mid_channels, stride, groups),
Conv1x1BNReLU(mid_channels, mid_channels)
)
if self.stride > 1:
self.shortcut = nn.Sequential(
Conv3x3BN(in_channels=in_channels, out_channels=in_channels, stride=stride, groups=groups),
Conv1x1BNReLU(in_channels, in_channels)
)
self.channel_shuffle = ChannelShuffle(groups)
def forward(self, x):
if self.stride > 1:
x1 = self.bottleneck(x)
x2 = self.shortcut(x)
else:
x1 = self.first_half(x)
x2 = self.second_split(x)
x1 = self.bottleneck(x1)
out = torch.cat([x1, x2], dim=1)
out = self.channel_shuffle(out)
return out
class ShuffleNetV2(nn.Module):
def __init__(self, planes, layers, groups, num_classes=1000):
super(ShuffleNetV2, self).__init__()
self.groups = groups
self.stage1 = nn.Sequential(
Conv3x3BNReLU(in_channels=3, out_channels=24, stride=2, groups=1),
nn.MaxPool2d(kernel_size=3, stride=2, padding=1)
)
self.stage2 = self._make_layer(24, planes[0], layers[0], True)
self.stage3 = self._make_layer(planes[0], planes[1], layers[1], False)
self.stage4 = self._make_layer(planes[1], planes[2], layers[2], False)
self.global_pool = nn.AdaptiveAvgPool2d(1)
self.dropout = nn.Dropout(p=0.2)
self.linear = nn.Linear(in_features=planes[2], out_features=num_classes)
self.init_params()
def _make_layer(self, in_channels, out_channels, block_num, is_stage2):
layers = []
layers.append(ShuffleNetUnits(in_channels=in_channels, out_channels=out_channels, stride= 2, groups=1 if is_stage2 else self.groups))
for idx in range(1, block_num):
layers.append(ShuffleNetUnits(in_channels=out_channels, out_channels=out_channels, stride=1, groups=self.groups))
return nn.Sequential(*layers)
def init_params(self):
for m in self.modules():
if isinstance(m, nn.Conv2d):
nn.init.kaiming_normal_(m.weight)
nn.init.constant_(m.bias, 0)
elif isinstance(m, nn.BatchNorm2d) or isinstance(m, nn.Linear):
nn.init.constant_(m.weight, 1)
nn.init.constant_(m.bias, 0)
def forward(self, x):
x = self.stage1(x)
x = self.stage2(x)
x = self.stage3(x)
x = self.stage4(x)
x = self.global_pool(x)
x = x.view(x.size(0), -1)
x = self.dropout(x)
out = self.linear(x)
return out
def shufflenet_v2_x2_0(**kwargs):
planes = [244, 488, 976]
layers = [4, 8, 4]
model = ShuffleNetV2(planes, layers, 1)
return model
def shufflenet_v2_x1_5(**kwargs):
planes = [176, 352, 704]
layers = [4, 8, 4]
model = ShuffleNetV2(planes, layers, 1)
return model
def shufflenet_v2_x1_0(**kwargs):
planes = [116, 232, 464]
layers = [4, 8, 4]
model = ShuffleNetV2(planes, layers, 1)
return model
def shufflenet_v2_x0_5(**kwargs):
planes = [48, 96, 192]
layers = [4, 8, 4]
model = ShuffleNetV2(planes, layers, 1)
return model
if __name__ == '__main__':
model = shufflenet_v2_x2_0()
print(model)
input = torch.randn(1, 3, 224, 224)
out = model(input)
print(out.shape)
