gan_cls_768.py

import torch
import torch.nn as nn
from torch.autograd import Variable
import numpy as np
import pdb

from torch.nn import functional as F
from torch.nn import init


'''

'''


class Concat_embed4(nn.Module):

    def __init__(self, embed_dim, projected_embed_dim):
        super(Concat_embed4, self).__init__()
        self.projection = nn.Sequential(
            nn.Linear(in_features=embed_dim, out_features=embed_dim),
            nn.BatchNorm1d(num_features=embed_dim),
            nn.LeakyReLU(negative_slope=0.2, inplace=True),

            nn.Linear(in_features=embed_dim, out_features=embed_dim),
            nn.BatchNorm1d(num_features=embed_dim),
            nn.LeakyReLU(negative_slope=0.2, inplace=True),

            nn.Linear(in_features=embed_dim, out_features=projected_embed_dim),
            nn.LeakyReLU(negative_slope=0.2, inplace=True),
        )

    def forward(self, inp, embed):
        projected_embed = self.projection(embed)
        replicated_embed = projected_embed.repeat(4, 4, 1, 1).permute(2,  3, 0, 1)
        hidden_concat = torch.cat([inp, replicated_embed], 1)
        return hidden_concat


class generator(nn.Module):
    def __init__(self):
        super(generator, self).__init__()
        self.image_size = 64
        self.num_channels = 3
        self.noise_dim = 100
        self.embed_dim = 768
        self.projected_embed_dim = 128
        self.latent_dim = self.noise_dim + self.projected_embed_dim
        self.ngf = 64

        self.projection = nn.Sequential(
            nn.Linear(in_features=self.embed_dim, out_features=self.embed_dim),
            nn.BatchNorm1d(num_features=self.embed_dim),
            nn.LeakyReLU(negative_slope=0.2, inplace=True),

            nn.Linear(in_features=self.embed_dim, out_features=self.embed_dim),
            nn.BatchNorm1d(num_features=self.embed_dim),
            nn.LeakyReLU(negative_slope=0.2, inplace=True),

            nn.Linear(in_features=self.embed_dim, out_features=self.projected_embed_dim),
            nn.BatchNorm1d(num_features=self.projected_embed_dim),
            nn.LeakyReLU(negative_slope=0.2, inplace=True)
        )

        self.netG = nn.ModuleList([
            nn.ConvTranspose2d(self.latent_dim, self.ngf * 8, 4, 1, 0, bias=False),
            nn.BatchNorm2d(self.ngf * 8),
            nn.ReLU(True),


            # state size. (ngf*8) x 4 x 4
            nn.ConvTranspose2d(self.ngf * 8, self.ngf * 4, 4, 2, 1, bias=False),
            nn.BatchNorm2d(self.ngf * 4),
            nn.ReLU(True),

            # state size. (ngf*4) x 8 x 8
            nn.ConvTranspose2d(self.ngf * 4, self.ngf * 2, 4, 2, 1, bias=False),
            nn.BatchNorm2d(self.ngf * 2),
            nn.ReLU(True),

            # state size. (ngf*2) x 16 x 16
            nn.ConvTranspose2d(self.ngf * 2, self.ngf, 4, 2, 1, bias=False),
            nn.BatchNorm2d(self.ngf),
            nn.ReLU(True),

            # state size. (ngf) x 32 x 32
            nn.ConvTranspose2d(self.ngf, self.num_channels, 4, 2, 1, bias=False),
            nn.Tanh()
            # state size. (num_channels) x 64 x 64
        ])

    def forward(self, embed_vector, z):
        projected_embed = self.projection(embed_vector)
        out = torch.cat([projected_embed.unsqueeze(2).unsqueeze(3), z], 1)
        for m in self.netG:
            out = m(out)
        return out


class discriminator(nn.Module):
    def __init__(self):
        super(discriminator, self).__init__()
        self.image_size = 64
        self.num_channels = 3
        self.embed_dim = 768
        self.projected_embed_dim = 128
        self.ndf = 64
        self.B_dim = 128
        self.C_dim = 16

        self.netD_1 = nn.Sequential(
            # input is (nc) x 64 x 64
            nn.Conv2d(self.num_channels, self.ndf, 4, 2, 1, bias=False),
            nn.LeakyReLU(0.2, inplace=True),
            # state size. (ndf) x 32 x 32

            # SelfAttention(self.ndf),
            nn.Conv2d(self.ndf, self.ndf * 2, 4, 2, 1, bias=False),
            nn.BatchNorm2d(self.ndf * 2),
            nn.LeakyReLU(0.2, inplace=True),

            # state size. (ndf*2) x 16 x 16

            nn.Conv2d(self.ndf * 2, self.ndf * 4, 4, 2, 1, bias=False),
            nn.BatchNorm2d(self.ndf * 4),
            nn.LeakyReLU(0.2, inplace=True),

            # state size. (ndf*4) x 8 x 8
            nn.Conv2d(self.ndf * 4, self.ndf * 8, 4, 2, 1, bias=False),
            nn.BatchNorm2d(self.ndf * 8),
            nn.LeakyReLU(0.2, inplace=True),
        )

        self.projector = Concat_embed4(self.embed_dim, self.projected_embed_dim)

        self.netD_2 = nn.Sequential(
            # state size. (ndf*8) x 4 x 4
            nn.Conv2d(self.ndf * 8 + self.projected_embed_dim,
                      self.ndf * 8, 1, 1, 0, bias=False),
            nn.BatchNorm2d(self.ndf * 8),
            nn.LeakyReLU(0.2, inplace=True),
            nn.Conv2d(self.ndf * 8, 1, 4, 1, 0, bias=False),
            nn.Sigmoid()
        )

    def forward(self, inp, embed):
        x_intermediate = self.netD_1(inp)
        x = self.projector(x_intermediate, embed)
        x = self.netD_2(x)

        return x.view(-1, 1).squeeze(1), x_intermediate