ccgan.py

import argparse
import os
import numpy as np
import math

import torchvision.transforms as transforms
from torchvision.utils import save_image

from torch.utils.data import DataLoader
from torchvision import datasets
from torch.autograd import Variable

from models import *

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

os.makedirs('images', exist_ok=True)

parser = argparse.ArgumentParser()
parser.add_argument('--n_epochs', type=int, default=200, help='number of epochs of training')
parser.add_argument('--batch_size', type=int, default=64, help='size of the batches')
parser.add_argument('--lr', type=float, default=0.0002, help='adam: learning rate')
parser.add_argument('--b1', type=float, default=0.5, help='adam: decay of first order momentum of gradient')
parser.add_argument('--b2', type=float, default=0.999, help='adam: decay of first order momentum of gradient')
parser.add_argument('--n_cpu', type=int, default=8, help='number of cpu threads to use during batch generation')
parser.add_argument('--latent_dim', type=int, default=100, help='dimensionality of the latent space')
parser.add_argument('--img_size', type=int, default=32, help='size of each image dimension')
parser.add_argument('--mask_size', type=int, default=8, help='size of random mask')
parser.add_argument('--channels', type=int, default=1, help='number of image channels')
parser.add_argument('--sample_interval', type=int, default=500, help='interval between image sampling')
opt = parser.parse_args()
print(opt)

cuda = True if torch.cuda.is_available() else False

# Calculate output of image discriminator (PatchGAN)
patch_h, patch_w = int(opt.img_size / 2**3), int(opt.img_size / 2**3)
patch = (1, patch_h, patch_w)

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

# Loss function
adversarial_loss = torch.nn.MSELoss()

# Initialize generator and discriminator
generator = Generator()
discriminator = Discriminator()

if cuda:
    generator.cuda()
    discriminator.cuda()
    adversarial_loss.cuda()

# Initialize weights
generator.apply(weights_init_normal)
discriminator.apply(weights_init_normal)

# Configure data loader
os.makedirs('../../data/cifar10', exist_ok=True)
cifar_loader = torch.utils.data.DataLoader(
    datasets.CIFAR10('../../data/cifar10', train=True, download=True,
                   transform=transforms.Compose([
                       transforms.Resize(opt.img_size),
                       transforms.ToTensor(),
                       transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5))
                   ])),
    batch_size=opt.batch_size, shuffle=True)

# Optimizers
optimizer_G = torch.optim.Adam(generator.parameters(), lr=opt.lr, betas=(opt.b1, opt.b2))
optimizer_D = torch.optim.Adam(discriminator.parameters(), lr=opt.lr, betas=(opt.b1, opt.b2))

Tensor = torch.cuda.FloatTensor if cuda else torch.FloatTensor

# Adversarial ground truths
valid = Variable(Tensor(np.ones(patch)), requires_grad=False)
fake = Variable(Tensor(np.zeros(patch)), requires_grad=False)

def apply_random_mask(imgs):
    idx = np.random.randint(0, opt.img_size-opt.mask_size, (imgs.shape[0], 2))

    for i, (y1, x1) in enumerate(idx):
        y2, x2 = y1 + opt.mask_size, x1 + opt.mask_size
        imgs[i, :, y1:y2, x1:x2] = -1

    return imgs

for epoch in range(opt.n_epochs):
    for i, (imgs, _) in enumerate(cifar_loader):

        masked_imgs = apply_random_mask(imgs)

        # Adversarial ground truths
        valid = Variable(Tensor(imgs.shape[0], *patch).fill_(1.0), requires_grad=False)
        fake = Variable(Tensor(imgs.shape[0], *patch).fill_(0.0), requires_grad=False)

        if cuda:
            imgs = imgs.type(torch.cuda.FloatTensor)
            masked_imgs = masked_imgs.type(torch.cuda.FloatTensor)

        real_imgs = Variable(imgs)
        masked_imgs = Variable(masked_imgs)

        # -----------------
        #  Train Generator
        # -----------------

        optimizer_G.zero_grad()

        # Generate a batch of images
        gen_imgs = generator(masked_imgs)

        # Loss measures generator's ability to fool the discriminator
        g_loss = adversarial_loss(discriminator(gen_imgs), valid)

        g_loss.backward()
        optimizer_G.step()

        # ---------------------
        #  Train Discriminator
        # ---------------------

        optimizer_D.zero_grad()

        # Measure discriminator's ability to classify real from generated samples
        real_loss = adversarial_loss(discriminator(real_imgs), valid)
        fake_loss = adversarial_loss(discriminator(gen_imgs.detach()), fake)
        d_loss = 0.5 * (real_loss + fake_loss)

        d_loss.backward()
        optimizer_D.step()

        print ("[Epoch %d/%d] [Batch %d/%d] [D loss: %f] [G loss: %f]" % (epoch, opt.n_epochs, i, len(cifar_loader),
                                                            d_loss.data.cpu().numpy()[0], g_loss.data.cpu().numpy()[0]))

        batches_done = epoch * len(dataloader) + i
        if batches_done % opt.sample_interval == 0:
            save_image(torch.cat((masked_imgs.data[:5], gen_imgs.data[:5], real_imgs.data[:5]), -2),
                        'images/%d.png' % batches_done, nrow=5, normalize=True)