se_inception_v3.py

import os
import torch
import torch.nn as nn
import torch.nn.functional as F
import torch.utils.model_zoo as model_zoo
from collections import defaultdict
from tqdm import tqdm
from torch.autograd import Variable
from utils import get_state_dict
from label_id_dict import label_to_category_id


class SELayer(nn.Module):
    def __init__(self, channel, reduction=16):
        super(SELayer, self).__init__()
        self.avg_pool = nn.AdaptiveAvgPool2d(1)
        self.fc = nn.Sequential(
                nn.Linear(channel, reduction),
                nn.ReLU(inplace=True),
                nn.Linear(reduction, channel),
                nn.Sigmoid()
        )
        self.init_parameters()

    def forward(self, x):
        b, c, _, _ = x.size()
        y = self.avg_pool(x).view(b, c)
        y = self.fc(y).view(b, c, 1, 1)
        return x * y

    def init_parameters(self):
        self.fc._modules['0'].weight.data.normal_(0, 1)
        self.fc._modules['0'].bias.data.zero_()
        self.fc._modules['2'].weight.data.normal_(0, 1)
        self.fc._modules['2'].bias.data.zero_()


class BasicConv2d(nn.Module):
    def __init__(self, in_channels, out_channels, **kwargs):
        super(BasicConv2d, self).__init__()
        self.conv = nn.Conv2d(in_channels, out_channels, bias=False, **kwargs)
        self.bn = nn.BatchNorm2d(out_channels, eps=0.001)

    def forward(self, x):
        x = self.conv(x)
        x = self.bn(x)
        return F.relu(x, inplace=True)


class SEInception3(nn.Module):
    use_cuda = torch.cuda.is_available()

    def __init__(self,  num_classes=1000):
        super(SEInception3, self).__init__()
        self.num_classes = num_classes
        self.name = "SEInception3"

        self.Conv2d_1a_3x3 = BasicConv2d(3, 32, kernel_size=3, stride=2)
        self.Conv2d_2a_3x3 = BasicConv2d(32, 32, kernel_size=3)
        self.Conv2d_2b_3x3 = BasicConv2d(32, 64, kernel_size=3, padding=1)
        self.Conv2d_3b_1x1 = BasicConv2d(64, 80, kernel_size=1)
        self.Conv2d_4a_3x3 = BasicConv2d(80, 192, kernel_size=3)
        self.Mixed_5b = InceptionA(192, pool_features=32)
        self.Mixed_5c = InceptionA(256, pool_features=64)
        self.Mixed_5d = InceptionA(288, pool_features=64)
        self.Mixed_6a = InceptionB(288)
        self.Mixed_6b = InceptionC(768, channels_7x7=128)
        self.Mixed_6c = InceptionC(768, channels_7x7=160)
        self.Mixed_6d = InceptionC(768, channels_7x7=160)
        self.Mixed_6e = InceptionC(768, channels_7x7=192)
        self.Mixed_7a = InceptionD(768)
        self.Mixed_7b = InceptionE(1280)
        self.Mixed_7c = InceptionE(2048)

        # apply 'Squeeze-and-Excitation'(SE) block
        self.Mixed_5b.add_module("SELayer", SELayer(192))
        self.Mixed_5c.add_module("SELayer", SELayer(256))
        self.Mixed_5d.add_module("SELayer", SELayer(288))
        self.Mixed_6a.add_module("SELayer", SELayer(288))
        self.Mixed_6b.add_module("SELayer", SELayer(768))
        self.Mixed_6c.add_module("SELayer", SELayer(768))
        self.Mixed_6d.add_module("SELayer", SELayer(768))
        self.Mixed_6e.add_module("SELayer", SELayer(768))
        self.Mixed_7a.add_module("SELayer", SELayer(768))
        self.Mixed_7b.add_module("SELayer", SELayer(1280))
        self.Mixed_7c.add_module("SELayer", SELayer(2048))

        self.fc = nn.Linear(2048, num_classes)

        # initialize the weight in SE blocks
        for m in self.modules():
            if isinstance(m, nn.Conv2d) or isinstance(m, nn.Linear):
                import scipy.stats as stats
                stddev = m.stddev if hasattr(m, 'stddev') else 0.1
                initial_weight = stats.truncnorm(-2, 2, scale=stddev)
                values = torch.Tensor(initial_weight.rvs(m.weight.data.numel()))
                values = values.view(m.weight.data.size())
                m.weight.data.copy_(values)
            elif isinstance(m, nn.BatchNorm2d):
                m.weight.data.fill_(1)
                m.bias.data.zero_()

    def forward(self, x):
        # 299 x 299 x 3
        x = self.Conv2d_1a_3x3(x)
        # 149 x 149 x 32
        x = self.Conv2d_2a_3x3(x)
        # 147 x 147 x 32
        x = self.Conv2d_2b_3x3(x)
        # 147 x 147 x 64
        x = F.max_pool2d(x, kernel_size=3, stride=2)
        # 73 x 73 x 64
        x = self.Conv2d_3b_1x1(x)
        # 73 x 73 x 80
        x = self.Conv2d_4a_3x3(x)
        # 71 x 71 x 192
        x = F.max_pool2d(x, kernel_size=3, stride=2)
        # 35 x 35 x 192
        x = self.Mixed_5b(x)
        # 35 x 35 x 256
        x = self.Mixed_5c(x)
        # 35 x 35 x 288
        x = self.Mixed_5d(x)
        # 35 x 35 x 288
        x = self.Mixed_6a(x)
        # 17 x 17 x 768
        x = self.Mixed_6b(x)
        # 17 x 17 x 768
        x = self.Mixed_6c(x)
        # 17 x 17 x 768
        x = self.Mixed_6d(x)
        # 17 x 17 x 768
        x = self.Mixed_6e(x)
        # 17 x 17 x 768
 
        # 17 x 17 x 768
        x = self.Mixed_7a(x)
        # 8 x 8 x 1280
        x = self.Mixed_7b(x)
        # 8 x 8 x 2048
        x = self.Mixed_7c(x)
        # 8 x 8 x 2048
 
        x = F.adaptive_avg_pool2d(x, output_size=1)

        # 1 x 1 x 2048
        x = F.dropout(x, training=self.training)
        # 1 x 1 x 2048
        x = x.view(x.size(0), -1)
        # 2048
        x = self.fc(x)
        # 1000 (num_classes)
 
        return x

    def load_pretrained_model(self, pretrained_model_file=None, skip=[]):
        if pretrained_model_file:
            pretrain_state_dict = get_state_dict(pretrained_model_file)
            state_dict = self.state_dict()
            keys = list(state_dict.keys())
            for key in keys:
                if any(s in key for s in skip):
                    continue
                try:
                    state_dict[key] = pretrain_state_dict[key]
                except KeyError:
                    print("KeyError: {} dosen't lie in pretrain state dict".format(key))
                    continue
        else:
            state_dict = model_zoo.load_url('https://download.pytorch.org/models/inception_v3_google-1a9a5a14.pth')
        self.load_state_dict(state_dict)
        pass

    def predict(self, data_loader, vote=False):
        if not vote:
            predicts = []
            for data in tqdm(data_loader):
                data = data.cuda() if self.use_cuda else data
                data = Variable(data, volatile=True)
                output = self.forward(data)
                category_label = output.data.max(1)[1]
                for i in range(len(category_label)):
                    predicts.append(label_to_category_id[category_label[i]])
        else:
            predicts = defaultdict(list)
            for item_id, data in tqdm(data_loader):
                data = data.cuda() if self.use_cuda else data
                data = Variable(data, volatile=True)
                output = self.forward(data)
                category_label = output.data.max(1)[1]
                for i in range(len(category_label)):
                    predicts[item_id[i]].append(label_to_category_id[category_label[i]])
        return predicts

    def save(self, file):
        torch.save(self.state_dict(), file)


class InceptionA(nn.Module):
    def __init__(self, in_channels, pool_features):
        super(InceptionA, self).__init__()
        self.branch1x1 = BasicConv2d(in_channels, 64, kernel_size=1)

        self.branch5x5_1 = BasicConv2d(in_channels, 48, kernel_size=1)
        self.branch5x5_2 = BasicConv2d(48, 64, kernel_size=5, padding=2)

        self.branch3x3dbl_1 = BasicConv2d(in_channels, 64, kernel_size=1)
        self.branch3x3dbl_2 = BasicConv2d(64, 96, kernel_size=3, padding=1)
        self.branch3x3dbl_3 = BasicConv2d(96, 96, kernel_size=3, padding=1)

        self.branch_pool = BasicConv2d(in_channels, pool_features, kernel_size=1)

    def forward(self, x):
        branch1x1 = self.branch1x1(x)

        branch5x5 = self.branch5x5_1(x)
        branch5x5 = self.branch5x5_2(branch5x5)

        branch3x3dbl = self.branch3x3dbl_1(x)
        branch3x3dbl = self.branch3x3dbl_2(branch3x3dbl)
        branch3x3dbl = self.branch3x3dbl_3(branch3x3dbl)

        branch_pool = F.avg_pool2d(x, kernel_size=3, stride=1, padding=1)
        branch_pool = self.branch_pool(branch_pool)

        outputs = [branch1x1, branch5x5, branch3x3dbl, branch_pool]
        return torch.cat(outputs, 1)


class InceptionB(nn.Module):
    def __init__(self, in_channels):
        super(InceptionB, self).__init__()
        self.branch3x3 = BasicConv2d(in_channels, 384, kernel_size=3, stride=2)

        self.branch3x3dbl_1 = BasicConv2d(in_channels, 64, kernel_size=1)
        self.branch3x3dbl_2 = BasicConv2d(64, 96, kernel_size=3, padding=1)
        self.branch3x3dbl_3 = BasicConv2d(96, 96, kernel_size=3, stride=2)

    def forward(self, x):
        branch3x3 = self.branch3x3(x)

        branch3x3dbl = self.branch3x3dbl_1(x)
        branch3x3dbl = self.branch3x3dbl_2(branch3x3dbl)
        branch3x3dbl = self.branch3x3dbl_3(branch3x3dbl)

        branch_pool = F.max_pool2d(x, kernel_size=3, stride=2)

        outputs = [branch3x3, branch3x3dbl, branch_pool]
        return torch.cat(outputs, 1)


class InceptionC(nn.Module):
    def __init__(self, in_channels, channels_7x7):
        super(InceptionC, self).__init__()
        self.branch1x1 = BasicConv2d(in_channels, 192, kernel_size=1)

        c7 = channels_7x7
        self.branch7x7_1 = BasicConv2d(in_channels, c7, kernel_size=1)
        self.branch7x7_2 = BasicConv2d(c7, c7, kernel_size=(1, 7), padding=(0, 3))
        self.branch7x7_3 = BasicConv2d(c7, 192, kernel_size=(7, 1), padding=(3, 0))

        self.branch7x7dbl_1 = BasicConv2d(in_channels, c7, kernel_size=1)
        self.branch7x7dbl_2 = BasicConv2d(c7, c7, kernel_size=(7, 1), padding=(3, 0))
        self.branch7x7dbl_3 = BasicConv2d(c7, c7, kernel_size=(1, 7), padding=(0, 3))
        self.branch7x7dbl_4 = BasicConv2d(c7, c7, kernel_size=(7, 1), padding=(3, 0))
        self.branch7x7dbl_5 = BasicConv2d(c7, 192, kernel_size=(1, 7), padding=(0, 3))

        self.branch_pool = BasicConv2d(in_channels, 192, kernel_size=1)

    def forward(self, x):
        branch1x1 = self.branch1x1(x)

        branch7x7 = self.branch7x7_1(x)
        branch7x7 = self.branch7x7_2(branch7x7)
        branch7x7 = self.branch7x7_3(branch7x7)

        branch7x7dbl = self.branch7x7dbl_1(x)
        branch7x7dbl = self.branch7x7dbl_2(branch7x7dbl)
        branch7x7dbl = self.branch7x7dbl_3(branch7x7dbl)
        branch7x7dbl = self.branch7x7dbl_4(branch7x7dbl)
        branch7x7dbl = self.branch7x7dbl_5(branch7x7dbl)

        branch_pool = F.avg_pool2d(x, kernel_size=3, stride=1, padding=1)
        branch_pool = self.branch_pool(branch_pool)

        outputs = [branch1x1, branch7x7, branch7x7dbl, branch_pool]
        return torch.cat(outputs, 1)


class InceptionD(nn.Module):
    def __init__(self, in_channels):
        super(InceptionD, self).__init__()
        self.branch3x3_1 = BasicConv2d(in_channels, 192, kernel_size=1)
        self.branch3x3_2 = BasicConv2d(192, 320, kernel_size=3, stride=2)

        self.branch7x7x3_1 = BasicConv2d(in_channels, 192, kernel_size=1)
        self.branch7x7x3_2 = BasicConv2d(192, 192, kernel_size=(1, 7), padding=(0, 3))
        self.branch7x7x3_3 = BasicConv2d(192, 192, kernel_size=(7, 1), padding=(3, 0))
        self.branch7x7x3_4 = BasicConv2d(192, 192, kernel_size=3, stride=2)

    def forward(self, x):
        branch3x3 = self.branch3x3_1(x)
        branch3x3 = self.branch3x3_2(branch3x3)

        branch7x7x3 = self.branch7x7x3_1(x)
        branch7x7x3 = self.branch7x7x3_2(branch7x7x3)
        branch7x7x3 = self.branch7x7x3_3(branch7x7x3)
        branch7x7x3 = self.branch7x7x3_4(branch7x7x3)

        branch_pool = F.max_pool2d(x, kernel_size=3, stride=2)
        outputs = [branch3x3, branch7x7x3, branch_pool]
        return torch.cat(outputs, 1)


class InceptionE(nn.Module):
    def __init__(self, in_channels):
        super(InceptionE, self).__init__()
        self.branch1x1 = BasicConv2d(in_channels, 320, kernel_size=1)

        self.branch3x3_1 = BasicConv2d(in_channels, 384, kernel_size=1)
        self.branch3x3_2a = BasicConv2d(384, 384, kernel_size=(1, 3), padding=(0, 1))
        self.branch3x3_2b = BasicConv2d(384, 384, kernel_size=(3, 1), padding=(1, 0))

        self.branch3x3dbl_1 = BasicConv2d(in_channels, 448, kernel_size=1)
        self.branch3x3dbl_2 = BasicConv2d(448, 384, kernel_size=3, padding=1)
        self.branch3x3dbl_3a = BasicConv2d(384, 384, kernel_size=(1, 3), padding=(0, 1))
        self.branch3x3dbl_3b = BasicConv2d(384, 384, kernel_size=(3, 1), padding=(1, 0))

        self.branch_pool = BasicConv2d(in_channels, 192, kernel_size=1)

    def forward(self, x):
        branch1x1 = self.branch1x1(x)

        branch3x3 = self.branch3x3_1(x)
        branch3x3 = [
            self.branch3x3_2a(branch3x3),
            self.branch3x3_2b(branch3x3),
        ]
        branch3x3 = torch.cat(branch3x3, 1)

        branch3x3dbl = self.branch3x3dbl_1(x)
        branch3x3dbl = self.branch3x3dbl_2(branch3x3dbl)
        branch3x3dbl = [
            self.branch3x3dbl_3a(branch3x3dbl),
            self.branch3x3dbl_3b(branch3x3dbl),
        ]
        branch3x3dbl = torch.cat(branch3x3dbl, 1)

        branch_pool = F.avg_pool2d(x, kernel_size=3, stride=1, padding=1)
        branch_pool = self.branch_pool(branch_pool)

        outputs = [branch1x1, branch3x3, branch3x3dbl, branch_pool]
        return torch.cat(outputs, 1)


class InceptionAux(nn.Module):
    def __init__(self, in_channels, num_classes):
        super(InceptionAux, self).__init__()
        self.conv0 = BasicConv2d(in_channels, 128, kernel_size=1)
        self.conv1 = BasicConv2d(128, 768, kernel_size=5)
        self.conv1.stddev = 0.01
        self.fc = nn.Linear(768, num_classes)
        self.fc.stddev = 0.001

    def forward(self, x):
        # 17 x 17 x 768
        x = F.avg_pool2d(x, kernel_size=5, stride=3)
        # 5 x 5 x 768
        x = self.conv0(x)
        # 5 x 5 x 128
        x = self.conv1(x)
        # 1 x 1 x 768
        x = x.view(x.size(0), -1)
        # 768
        x = self.fc(x)
        # 1000
        return x


if __name__ == '__main__':

    def test_net():
        batch_size = 10
        num_classes = 2000
        C, H, W = 3, 150, 150

        inputs = torch.randn(batch_size, C, H, W)
        labels = torch.randn(batch_size, num_classes)
        in_shape = inputs.size()[1:]

        net = SEInception3(num_classes=num_classes)
        net.load_pretrained_model(
            'saved_models/inception_v3_google-1a9a5a14.pth',
            skip=['fc.weight', 'fc.bias']
        )
        net.train()
        print(type(net))
        # print(net)

        from torch.optim import SGD
        optimizer = SGD(filter(lambda p: p.requires_grad, net.parameters()), lr=0.01, momentum=0.9, weight_decay=0.0001)
        optimizer.zero_grad()

        print("Begin Training!")
        x = Variable(inputs)
        y = Variable(labels)
        logits = net.forward(x)
        loss = F.cross_entropy(logits, y)
        loss.backward()
        optimizer.step()

        print("Begin Evaluation!")
        net.eval()
        inputs = torch.randn(1, C, H, W)
        inputs = Variable(inputs)
        probs = F.softmax(net.forward(inputs))
        print(torch.max(probs))

    print( '%s: calling main function ... ' % os.path.basename(__file__))
    test_net()