deep-person-reid/torchreid/models/squeezenet.py

from __future__ import absolute_import
from __future__ import division

from collections import OrderedDict
import math

import torch
import torch.nn as nn
from torch.utils import model_zoo
from torch.nn import functional as F
import torch.nn.init as init
import torchvision
import torch.utils.model_zoo as model_zoo


__all__ = ['squeezenet1_0', 'squeezenet1_1']


model_urls = {
    'squeezenet1_0': 'https://download.pytorch.org/models/squeezenet1_0-a815701f.pth',
    'squeezenet1_1': 'https://download.pytorch.org/models/squeezenet1_1-f364aa15.pth',
}


class Fire(nn.Module):

    def __init__(self, inplanes, squeeze_planes,
                 expand1x1_planes, expand3x3_planes):
        super(Fire, self).__init__()
        self.inplanes = inplanes
        self.squeeze = nn.Conv2d(inplanes, squeeze_planes, kernel_size=1)
        self.squeeze_activation = nn.ReLU(inplace=True)
        self.expand1x1 = nn.Conv2d(squeeze_planes, expand1x1_planes,
                                   kernel_size=1)
        self.expand1x1_activation = nn.ReLU(inplace=True)
        self.expand3x3 = nn.Conv2d(squeeze_planes, expand3x3_planes,
                                   kernel_size=3, padding=1)
        self.expand3x3_activation = nn.ReLU(inplace=True)

    def forward(self, x):
        x = self.squeeze_activation(self.squeeze(x))
        return torch.cat([
            self.expand1x1_activation(self.expand1x1(x)),
            self.expand3x3_activation(self.expand3x3(x))
        ], 1)


class SqueezeNet(nn.Module):
    """
    SqueezeNet

    Reference:
    Iandola et al. SqueezeNet: AlexNet-level accuracy with 50x fewer parameters
    and< 0.5 MB model size. arXiv:1602.07360.
    """
    def __init__(self, num_classes, loss, version=1.0, **kwargs):
        super(SqueezeNet, self).__init__()
        self.loss = loss
        if version not in [1.0, 1.1]:
            raise ValueError("Unsupported SqueezeNet version {version}:"
                             "1.0 or 1.1 expected".format(version=version))
        self.num_classes = num_classes
        if version == 1.0:
            self.features = nn.Sequential(
                nn.Conv2d(3, 96, kernel_size=7, stride=2),
                nn.ReLU(inplace=True),
                nn.MaxPool2d(kernel_size=3, stride=2, ceil_mode=True),
                Fire(96, 16, 64, 64),
                Fire(128, 16, 64, 64),
                Fire(128, 32, 128, 128),
                nn.MaxPool2d(kernel_size=3, stride=2, ceil_mode=True),
                Fire(256, 32, 128, 128),
                Fire(256, 48, 192, 192),
                Fire(384, 48, 192, 192),
                Fire(384, 64, 256, 256),
                nn.MaxPool2d(kernel_size=3, stride=2, ceil_mode=True),
                Fire(512, 64, 256, 256),
            )
        else:
            self.features = nn.Sequential(
                nn.Conv2d(3, 64, kernel_size=3, stride=2),
                nn.ReLU(inplace=True),
                nn.MaxPool2d(kernel_size=3, stride=2, ceil_mode=True),
                Fire(64, 16, 64, 64),
                Fire(128, 16, 64, 64),
                nn.MaxPool2d(kernel_size=3, stride=2, ceil_mode=True),
                Fire(128, 32, 128, 128),
                Fire(256, 32, 128, 128),
                nn.MaxPool2d(kernel_size=3, stride=2, ceil_mode=True),
                Fire(256, 48, 192, 192),
                Fire(384, 48, 192, 192),
                Fire(384, 64, 256, 256),
                Fire(512, 64, 256, 256),
            )
        # Final convolution is initialized differently form the rest
        final_conv = nn.Conv2d(512, self.num_classes, kernel_size=1)
        self.classifier = nn.Sequential(
            nn.Dropout(p=0.5),
            final_conv,
            nn.ReLU(inplace=True),
            nn.AdaptiveAvgPool2d(1)
        )

        for m in self.modules():
            if isinstance(m, nn.Conv2d):
                if m is final_conv:
                    init.normal_(m.weight, mean=0.0, std=0.01)
                else:
                    init.kaiming_uniform_(m.weight)
                if m.bias is not None:
                    init.constant_(m.bias, 0)

    def forward(self, x):
        f = self.features(x)

        if not self.training:
            v = F.adaptive_avg_pool2d(f, 1)
            v = v.view(v.size(0), -1)
            return v

        y = self.classifier(f)
        y = y.view(y.size(0), self.num_classes)

        if self.loss == {'xent'}:
            return y
        elif self.loss == {'xent', 'htri'}:
            v = F.adaptive_avg_pool2d(f, 1)
            v = v.view(v.size(0), -1)
            return y, v
        else:
            raise KeyError("Unsupported loss: {}".format(self.loss))


def init_pretrained_weights(model, model_url):
    """
    Initialize model with pretrained weights.
    Layers that don't match with pretrained layers in name or size are kept unchanged.
    """
    pretrain_dict = model_zoo.load_url(model_url, map_location=None)
    model_dict = model.state_dict()
    pretrain_dict = {k: v for k, v in pretrain_dict.items() if k in model_dict and model_dict[k].size() == v.size()}
    model_dict.update(pretrain_dict)
    model.load_state_dict(model_dict)
    print("Initialized model with pretrained weights from {}".format(model_url))


def squeezenet1_0(num_classes, loss, pretrained=True, **kwargs):
    model = SqueezeNet(num_classes, loss, version=1.0, **kwargs)
    if pretrained:
        init_pretrained_weights(model, model_urls['squeezenet1_0'])
    return model


def squeezenet1_1(num_classes, loss, pretrained=True, **kwargs):
    model = SqueezeNet(num_classes, loss, version=1.1, **kwargs)
    if pretrained:
        init_pretrained_weights(model, model_urls['squeezenet1_1'])
    return model