model.py

import torch
import torch.nn as nn
import torch.nn.functional as F
import numpy as np
import random

from PIL import Image


class BaseNetwork(nn.Module):
    def __init__(self):
        super(BaseNetwork, self).__init__()

    def init_weights(self):
        self.apply(self._weights_init_fn)

    def _weights_init_fn(self, m):
        classname = m.__class__.__name__
        if classname.find('Conv') != -1:
            m.weight.data.normal_(0.0, 0.02)
            if hasattr(m.bias, 'data'):
                m.bias.data.fill_(0)
        elif classname.find('BatchNorm2d') != -1:
            m.weight.data.normal_(1.0, 0.02)
            m.bias.data.fill_(0)


class ResidualBlock(BaseNetwork):
    """Residual Block with instance normalization."""

    def __init__(self, dim_in, dim_out):
        super(ResidualBlock, self).__init__()
        self.main = nn.Sequential(
            nn.Conv2d(dim_in, dim_out, kernel_size=3,
                      stride=1, padding=1, bias=False),
            nn.InstanceNorm2d(dim_out, affine=True, track_running_stats=True),
            nn.ReLU(inplace=True),
            nn.Conv2d(dim_out, dim_out, kernel_size=3,
                      stride=1, padding=1, bias=False),
            nn.InstanceNorm2d(dim_out, affine=True, track_running_stats=True))

        self.init_weights()

    def forward(self, x):
        return x + self.main(x)


class Generator(BaseNetwork):
    """Generator network."""

    def __init__(self, conv_dim=64, c_dim=5, repeat_num=6):
        super(Generator, self).__init__()

        layers = []
        layers.append(nn.Conv2d(3+c_dim, conv_dim, kernel_size=7,
                                stride=1, padding=3, bias=False))
        layers.append(nn.InstanceNorm2d(
            conv_dim, affine=True, track_running_stats=True))
        layers.append(nn.ReLU(inplace=True))

        self.debug1 = nn.Sequential(*layers)

        # Down-sampling layers.
        curr_dim = conv_dim
        for i in range(2):
            layers.append(nn.Conv2d(curr_dim, curr_dim*2,
                                    kernel_size=4, stride=2, padding=1, bias=False))
            layers.append(nn.InstanceNorm2d(
                curr_dim*2, affine=True, track_running_stats=True))
            layers.append(nn.ReLU(inplace=True))
            curr_dim = curr_dim * 2

        self.debug2 = nn.Sequential(*layers)

        # Bottleneck layers.
        for i in range(repeat_num):
            layers.append(ResidualBlock(dim_in=curr_dim, dim_out=curr_dim))

        self.debug3 = nn.Sequential(*layers)

        # Up-sampling layers.
        for i in range(2):
            layers.append(nn.ConvTranspose2d(curr_dim, curr_dim //
                                             2, kernel_size=4, stride=2, padding=1, bias=False))
            layers.append(nn.InstanceNorm2d(
                curr_dim//2, affine=True, track_running_stats=True))
            layers.append(nn.ReLU(inplace=True))
            curr_dim = curr_dim // 2

        self.main = nn.Sequential(*layers)

        self.debug4 = nn.Sequential(*layers)

        # Same architecture for the color regression
        layers = []
        layers.append(nn.Conv2d(curr_dim, 3, kernel_size=7,
                                stride=1, padding=3, bias=False))
        layers.append(nn.Tanh())
        self.im_reg = nn.Sequential(*layers)

        # One Channel output and Sigmoid function for the attention layer
        layers = []
        layers.append(nn.Conv2d(curr_dim, 1, kernel_size=7,
                                stride=1, padding=3, bias=False))
        layers.append(nn.Sigmoid())  # Values between 0 and 1
        self.im_att = nn.Sequential(*layers)

        self.init_weights()

    def forward(self, x, c):
        # Replicate spatially and concatenate domain information.

        c = c.unsqueeze(2).unsqueeze(3)
        c = c.expand(c.size(0), c.size(1), x.size(2), x.size(3))

        x = torch.cat([x, c], dim=1)
        features = self.main(x)

        reg = self.im_reg(features)
        att = self.im_att(features)

        return att, reg


class Discriminator(BaseNetwork):
    """Discriminator network with PatchGAN."""

    def __init__(self, image_size=128, conv_dim=64, c_dim=5, repeat_num=6):
        super(Discriminator, self).__init__()

        layers = []
        layers.append(
            nn.Conv2d(3, conv_dim, kernel_size=4, stride=2, padding=1))
        layers.append(nn.LeakyReLU(0.01))

        curr_dim = conv_dim
        for i in range(1, repeat_num):
            layers.append(nn.Conv2d(curr_dim, curr_dim*2,
                                    kernel_size=4, stride=2, padding=1))
            layers.append(nn.LeakyReLU(0.01))
            curr_dim = curr_dim * 2

        kernel_size = int(image_size / np.power(2, repeat_num))
        self.main = nn.Sequential(*layers)
        self.conv1 = nn.Conv2d(curr_dim, 1, kernel_size=3,
                               stride=1, padding=1, bias=False)
        self.conv2 = nn.Conv2d(
            curr_dim, c_dim, kernel_size=kernel_size, bias=False)

        self.init_weights()

    def forward(self, x):
        h = self.main(x)
        out_src = self.conv1(h)
        out_cls = self.conv2(h)

        # out_cls.view(out_cls.size(0), out_cls.size(1))
        return out_src.squeeze(), out_cls.squeeze()