import math
import torch
import torch.nn as nn
import torch.nn.functional as F
[docs]class BasicBlock(nn.Module):
[docs] def __init__(self, in_planes, out_planes, stride, dropRate=0.0):
super(BasicBlock, self).__init__()
self.bn1 = nn.BatchNorm2d(in_planes)
self.relu1 = nn.ReLU(inplace=True)
self.conv1 = nn.Conv2d(in_planes, out_planes, kernel_size=3, stride=stride,
padding=1, bias=False)
self.bn2 = nn.BatchNorm2d(out_planes)
self.relu2 = nn.ReLU(inplace=True)
self.conv2 = nn.Conv2d(out_planes, out_planes, kernel_size=3, stride=1,
padding=1, bias=False)
self.droprate = dropRate
self.equalInOut = (in_planes == out_planes)
self.convShortcut = (not self.equalInOut) and nn.Conv2d(in_planes, out_planes, kernel_size=1, stride=stride,
padding=0, bias=False) or None
[docs] def forward(self, x):
if not self.equalInOut:
x = self.relu1(self.bn1(x))
else:
out = self.relu1(self.bn1(x))
out = self.relu2(self.bn2(self.conv1(out if self.equalInOut else x)))
if self.droprate > 0:
out = F.dropout(out, p=self.droprate, training=self.training)
out = self.conv2(out)
return torch.add(x if self.equalInOut else self.convShortcut(x), out)
[docs]class NetworkBlock(nn.Module):
[docs] def __init__(self, nb_layers, in_planes, out_planes, block, stride, dropRate=0.0):
super(NetworkBlock, self).__init__()
self.layer = self._make_layer(block, in_planes, out_planes, nb_layers, stride, dropRate)
def _make_layer(self, block, in_planes, out_planes, nb_layers, stride, dropRate):
layers = []
for i in range(int(nb_layers)):
layers.append(block(i == 0 and in_planes or out_planes, out_planes, i == 0 and stride or 1, dropRate))
return nn.Sequential(*layers)
[docs] def forward(self, x):
return self.layer(x)
[docs]class WideResNet(nn.Module):
[docs] def __init__(self, depth=34, num_classes=10, widen_factor=1, dropRate=0.0, use_FNandWN=False):
super(WideResNet, self).__init__()
self.use_FNandWN = use_FNandWN
nChannels = [16, 16 * widen_factor, 32 * widen_factor, 64 * widen_factor]
assert ((depth - 4) % 6 == 0)
n = (depth - 4) // 6
block = BasicBlock
# 1st conv before any network block
self.conv1 = nn.Conv2d(3, nChannels[0], kernel_size=3, stride=1,
padding=1, bias=False)
# 1st block
self.block1 = NetworkBlock(n, nChannels[0], nChannels[1], block, 1, dropRate)
# 2nd block
self.block2 = NetworkBlock(n, nChannels[1], nChannels[2], block, 2, dropRate)
# 3rd block
self.block3 = NetworkBlock(n, nChannels[2], nChannels[3], block, 2, dropRate)
# global average pooling and classifier
self.bn1 = nn.BatchNorm2d(nChannels[3])
self.relu = nn.ReLU(inplace=True)
# self.fc = nn.Linear(nChannels[3], num_classes)
if self.use_FNandWN:
self.fc = nn.Linear(nChannels[3], num_classes, bias = False)
else:
self.fc = nn.Linear(nChannels[3], num_classes)
self.nChannels = nChannels[3]
for m in self.modules():
if isinstance(m, nn.Conv2d):
n = m.kernel_size[0] * m.kernel_size[1] * m.out_channels
m.weight.data.normal_(0, math.sqrt(2. / n))
elif isinstance(m, nn.BatchNorm2d):
m.weight.data.fill_(1)
m.bias.data.zero_()
elif isinstance(m, nn.Linear) and not self.use_FNandWN:
m.bias.data.zero_()
[docs] def forward(self, x):
out = self.conv1(x)
out = self.block1(out)
out = self.block2(out)
out = self.block3(out)
out = self.relu(self.bn1(out))
out = F.avg_pool2d(out, 8)
out = out.view(-1, self.nChannels)
if self.use_FNandWN:
out = F.normalize(out, p=2, dim=1)
for _, module in self.fc.named_modules():
if isinstance(module, nn.Linear):
module.weight.data = F.normalize(module.weight, p=2, dim=1)
return self.fc(out)
[docs]def create_wres34_10_fn():
'''The function to create wide-resnet34-10 with FN and WN for cifar10 models.'''
model = WideResNet(depth=34, num_classes=10, widen_factor=10, dropRate=0.0, use_FNandWN = True)
return model
[docs]def create_wres34_10():
'''The function to create wide-resnet34-10 for cifar10 models.'''
model = WideResNet(depth=34, num_classes=10, widen_factor=10, dropRate=0.0)
return model
[docs]def create_wres28_10():
'''The function to create wide-resnet28-10 for cifar10 models.'''
model = WideResNet(depth=28, num_classes=10, widen_factor=10, dropRate=0.0)
return model