'''
Reference:
https://github.com/khurramjaved96/incremental-learning/blob/autoencoders/model/resnet32.py
'''
import math
import torch
import torch.nn as nn
import torch.nn.functional as F
[docs]
class DownsampleA(nn.Module):
def __init__(self, nIn, nOut, stride):
super(DownsampleA, self).__init__()
assert stride == 2
self.avg = nn.AvgPool2d(kernel_size=1, stride=stride)
[docs]
def forward(self, x):
x = self.avg(x)
return torch.cat((x, x.mul(0)), 1)
[docs]
class DownsampleB(nn.Module):
def __init__(self, nIn, nOut, stride):
super(DownsampleB, self).__init__()
self.conv = nn.Conv2d(nIn, nOut, kernel_size=1, stride=stride, padding=0, bias=False)
self.bn = nn.BatchNorm2d(nOut)
[docs]
def forward(self, x):
x = self.conv(x)
x = self.bn(x)
return x
[docs]
class DownsampleC(nn.Module):
def __init__(self, nIn, nOut, stride):
super(DownsampleC, self).__init__()
assert stride != 1 or nIn != nOut
self.conv = nn.Conv2d(nIn, nOut, kernel_size=1, stride=stride, padding=0, bias=False)
[docs]
def forward(self, x):
x = self.conv(x)
return x
[docs]
class DownsampleD(nn.Module):
def __init__(self, nIn, nOut, stride):
super(DownsampleD, self).__init__()
assert stride == 2
self.conv = nn.Conv2d(nIn, nOut, kernel_size=2, stride=stride, padding=0, bias=False)
self.bn = nn.BatchNorm2d(nOut)
[docs]
def forward(self, x):
x = self.conv(x)
x = self.bn(x)
return x
[docs]
class ResNetBasicblock(nn.Module):
expansion = 1
def __init__(self, inplanes, planes, stride=1, downsample=None):
super(ResNetBasicblock, self).__init__()
self.conv_a = nn.Conv2d(inplanes, planes, kernel_size=3, stride=stride, padding=1, bias=False)
self.bn_a = nn.BatchNorm2d(planes)
self.conv_b = nn.Conv2d(planes, planes, kernel_size=3, stride=1, padding=1, bias=False)
self.bn_b = nn.BatchNorm2d(planes)
self.downsample = downsample
[docs]
def forward(self, x):
residual = x
basicblock = self.conv_a(x)
basicblock = self.bn_a(basicblock)
basicblock = F.relu(basicblock, inplace=True)
basicblock = self.conv_b(basicblock)
basicblock = self.bn_b(basicblock)
if self.downsample is not None:
residual = self.downsample(x)
return F.relu(residual + basicblock, inplace=True)
[docs]
class CifarResNet(nn.Module):
"""
ResNet optimized for the Cifar Dataset, as specified in
https://arxiv.org/abs/1512.03385.pdf
"""
def __init__(self, block, depth, channels=3):
super(CifarResNet, self).__init__()
# Model type specifies number of layers for CIFAR-10 and CIFAR-100 model
assert (depth - 2) % 6 == 0, 'depth should be one of 20, 32, 44, 56, 110'
layer_blocks = (depth - 2) // 6
self.conv_1_3x3 = nn.Conv2d(channels, 16, kernel_size=3, stride=1, padding=1, bias=False)
self.bn_1 = nn.BatchNorm2d(16)
self.inplanes = 16
self.stage_1 = self._make_layer(block, 16, layer_blocks, 1)
self.stage_2 = self._make_layer(block, 32, layer_blocks, 2)
self.stage_3 = self._make_layer(block, 64, layer_blocks, 2)
self.avgpool = nn.AvgPool2d(8)
self.out_dim = 64 * block.expansion
self.fc = nn.Linear(64 * block.expansion, 10)
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))
# m.bias.data.zero_()
elif isinstance(m, nn.BatchNorm2d):
m.weight.data.fill_(1)
m.bias.data.zero_()
elif isinstance(m, nn.Linear):
nn.init.kaiming_normal_(m.weight)
m.bias.data.zero_()
def _make_layer(self, block, planes, blocks, stride=1):
downsample = None
if stride != 1 or self.inplanes != planes * block.expansion:
downsample = DownsampleA(self.inplanes, planes * block.expansion, stride)
layers = []
layers.append(block(self.inplanes, planes, stride, downsample))
self.inplanes = planes * block.expansion
for i in range(1, blocks):
layers.append(block(self.inplanes, planes))
return nn.Sequential(*layers)
[docs]
def forward(self, x):
x = self.conv_1_3x3(x) # [bs, 16, 32, 32]
x = F.relu(self.bn_1(x), inplace=True)
x_1 = self.stage_1(x) # [bs, 16, 32, 32]
x_2 = self.stage_2(x_1) # [bs, 32, 16, 16]
x_3 = self.stage_3(x_2) # [bs, 64, 8, 8]
pooled = self.avgpool(x_3) # [bs, 64, 1, 1]
features = pooled.view(pooled.size(0), -1) # [bs, 64]
return {
'fmaps': [x_1, x_2, x_3],
'features': features
}
@property
def last_conv(self):
return self.stage_3[-1].conv_b
[docs]
def resnet20mnist():
"""Constructs a ResNet-20 model for MNIST."""
model = CifarResNet(ResNetBasicblock, 20, 1)
return model
[docs]
def resnet32mnist():
"""Constructs a ResNet-32 model for MNIST."""
model = CifarResNet(ResNetBasicblock, 32, 1)
return model
[docs]
def resnet20():
"""Constructs a ResNet-20 model for CIFAR-10."""
model = CifarResNet(ResNetBasicblock, 20)
return model
[docs]
def resnet32():
"""Constructs a ResNet-32 model for CIFAR-10."""
model = CifarResNet(ResNetBasicblock, 32)
return model
[docs]
def resnet44():
"""Constructs a ResNet-44 model for CIFAR-10."""
model = CifarResNet(ResNetBasicblock, 44)
return model
[docs]
def resnet56():
"""Constructs a ResNet-56 model for CIFAR-10."""
model = CifarResNet(ResNetBasicblock, 56)
return model
[docs]
def resnet110():
"""Constructs a ResNet-110 model for CIFAR-10."""
model = CifarResNet(ResNetBasicblock, 110)
return model