Dubhe/model_measuring/kamal/vision/models/classification/cifar/wrn.py

#Adapted from https://github.com/polo5/ZeroShotKnowledgeTransfer/blob/master/models/wresnet.py 

import math
import torch
import torch.nn as nn
import torch.nn.functional as F

__all__ = ['wrn']

class BasicBlock(nn.Module):
    def __init__(self, in_planes, out_planes, stride, dropout_rate=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.dropout = nn.Dropout( dropout_rate )
        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

    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)))
        out = self.dropout(out)
        out = self.conv2(out)
        return torch.add(x if self.equalInOut else self.convShortcut(x), out)


class NetworkBlock(nn.Module):
    def __init__(self, nb_layers, in_planes, out_planes, block, stride, dropout_rate=0.0):
        super(NetworkBlock, self).__init__()
        self.layer = self._make_layer(block, in_planes, out_planes, nb_layers, stride, dropout_rate)

    def _make_layer(self, block, in_planes, out_planes, nb_layers, stride, dropout_rate):
        layers = []
        for i in range(nb_layers):
            layers.append(block(i == 0 and in_planes or out_planes, out_planes, i == 0 and stride or 1, dropout_rate))
        return nn.Sequential(*layers)

    def forward(self, x):
        return self.layer(x)


class WideResNet(nn.Module):
    def __init__(self, depth, num_classes, widen_factor=1, dropout_rate=0.0):
        super(WideResNet, self).__init__()
        nChannels = [16, 16*widen_factor, 32*widen_factor, 64*widen_factor]
        assert (depth - 4) % 6 == 0, 'depth should be 6n+4'
        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, dropout_rate)
        # 2nd block
        self.block2 = NetworkBlock(n, nChannels[1], nChannels[2], block, 2, dropout_rate)
        # 3rd block
        self.block3 = NetworkBlock(n, nChannels[2], nChannels[3], block, 2, dropout_rate)
        # 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)
        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):
                m.bias.data.zero_()

    def forward(self, x, return_features=False):
        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)
        features = out.view(-1, self.nChannels)
        out = self.fc(features)
        if return_features:
            return out, features
        else:
            return out

def wrn_16_1(num_classes, dropout_rate=0):
    return WideResNet(depth=16, num_classes=num_classes, widen_factor=1, dropout_rate=dropout_rate)

def wrn_16_2(num_classes, dropout_rate=0):
    return WideResNet(depth=16, num_classes=num_classes, widen_factor=2, dropout_rate=dropout_rate)

def wrn_40_1(num_classes, dropout_rate=0):
    return WideResNet(depth=40, num_classes=num_classes, widen_factor=1, dropout_rate=dropout_rate)

def wrn_40_2(num_classes, dropout_rate=0):
    return WideResNet(depth=40, num_classes=num_classes, widen_factor=2, dropout_rate=dropout_rate)