PGGAN.py

#! python3
# -*- coding: utf-8 -*-
"""
This is followed by https://github.com/shanexn

################################################################################################
"""
import os
import random
import time
import mrcfile
import warnings
from PIL import Image
from model import pg_model
import torch.optim as optim
import torch.backends.cudnn as cudnn
import torchvision.utils as vutils
import numpy as np
import torch.utils.data as udata
import torchvision.datasets as vdatasets
import torchvision.transforms as transforms
import argparse

parser = argparse.ArgumentParser()

parser.add_argument('--save_path', type=str, default='WGANgp_l1')
parser.add_argument('--lamda', type= float, default = 100)
opt = parser.parse_args()

if not os.path.isdir('./PGGAN'):
    os.mkdir('PGGAN')
if not os.path.isdir('./PGGAN/' + opt.save_path):
    os.mkdir('./PGGAN/' + opt.save_path)
save_path_name = './PGGAN/' + opt.save_path
################################################################################
MANUAL_SEED = 0.5
CUDNN_BENCHMARK = True
TO_GPU = True  # For now, all run in GPU mode
DEVICE = torch.device("cuda:0")
G_LR = 0.0001
D_LR = 0.0001
ADAM_BETA = (0.0, 0.99)
ADAM_EPS = 1e-8
LOD_STABLE_IMGS = 2 * 1e5  # Thousands of real images to show before doubling the resolution.
LOD_FADEIN_IMGS = 2 * 1e5  # Thousands of real images to show when fading in new layers.
INIT_RESOLUTION = 4  # Image resolution used at the beginning training.
SAVE_IMG_STEP = 500
LAMBDA_FOR_WGANGP = 10
CRITIC_FOR_WGANGP = 1
MINIBATCH_MAPPING = {
    4: 16,
    8: 16,
    16: 16,
    32: 16,
    64: 16,
    128: 8,
    256: 4
}


#################################################################################
# Construct Generator and Discriminator #########################################
#################################################################################
class PGGAN(object):
    def __init__(self,
                 resolution,  # Resolution.
                 latent_size,  # Dimensionality of the latent vectors.
                 criterion_type="WGAN-GP",  # ["GAN", "WGAN-GP"]
                 minibatch_mapping=MINIBATCH_MAPPING,  # Batch size of each resolution
                 rgb_channel=1,  # Output channel size, for rgb always 1
                 fmap_base=2 ** 12,  # Overall multiplier for the number of feature maps.
                 fmap_decay=1.0,  # log2 feature map reduction when doubling the resolution.
                 fmap_max=2 ** 8,  # Maximum number of feature maps in any layer.
                 is_tanh=True,
                 is_sigmoid=True
                 ):
        self.latent_size_ = latent_size
        self.rgb_channel_ = rgb_channel
        self.fmap_base_ = fmap_base
        self.fmap_decay_ = fmap_decay
        self.fmap_max_ = fmap_max

        self.image_pyramid_ = int(np.log2(resolution))  # max level of the Image Pyramid
        self.resolution_ = 2 ** self.image_pyramid_  # correct resolution
        self.lod_init_res_ = INIT_RESOLUTION  # Image resolution used at the beginning.

        self.criterion_type_ = criterion_type
        self.is_tanh_ = is_tanh
        self.is_sigmoid_ = False if self.criterion_type_ in ["WGAN-GP"] else is_sigmoid

        self.gradient_weight_real_ = torch.tensor(-1.0).cuda()
        self.gradient_weight_fake_ = torch.tensor(1.0).cuda()

        # Mentioned on page13 of the Paper (Appendix A.1)
        # 4x4 - 128x128 -> 16
        # 256x256       -> 14
        # 512x512       -> 6
        # 1024x1024     -> 3
        # self.minibatch_mapping_ = {
        #     4: 16, 8: 16, 16: 16, 32: 16, 64: 16, 128: 16,
        #     256: 16, 512: 6, 1024: 3}
        self.minibatch_mapping_ = minibatch_mapping

        self._init_network()

    def gen_image_by_existed_models(self, model_path, file_name, fake_seed=None):

        import re
        regx = "(?<=x)\d+"

        if isinstance(model_path, (tuple, list)):
            g_model_path, d_model_path = model_path
        else:
            g_model_path = model_path
        if not os.path.exists(g_model_path):
            raise FileExistsError("Models not exist: {}".format(model_path))

        res = re.findall(regx, g_model_path)
        if res is None:
            raise TypeError("Models is not inappropriate: {}".format(model_path))
        res = int(res[0])
        # batch_size = self.minibatch_mapping_[res]
        batch_size = 1
        self.g_net.net_config = [int(np.log2(res)) - 2, "stable", 1.0]
        self.g_net.load_state_dict(torch.load(g_model_path))
        # self.d_net.load_state_dict(torch.load(d_model_path))
        if fake_seed is None:
            fake_seed = torch.randn(batch_size,
                                    self.latent_size_, 1, 1)
        fake_work = self.g_net(fake_seed)
        vutils.save_image(fake_work.detach(),
                          file_name, nrow=4, normalize=True, padding=0)

    def train(self, dataset_root):
        # init settings
        random.seed = MANUAL_SEED
        torch.manual_seed(MANUAL_SEED)
        cudnn.benchmark = CUDNN_BENCHMARK
        # self._init_network()
        self._init_optimizer()
        self._init_criterion()

        net_level = 0
        net_status = "stable"
        net_alpha = 1.0
        pyramid_from = int(np.log2(self.lod_init_res_))
        pyramid_to = self.image_pyramid_
        for cursor_pyramid in range(pyramid_from - 1, pyramid_to):
            print(cursor_pyramid)
            minibatch = self.minibatch_mapping_[2 ** (
                    cursor_pyramid + 1)]  # if not in the map, raise error directly  ############################################
            net_level = cursor_pyramid - 1
            stable_steps = int(LOD_STABLE_IMGS // minibatch)
            fadein_steps = int(LOD_FADEIN_IMGS // minibatch)

            current_level_res = 2 ** (net_level + 2)

            noisy_data = mrcfile.open(
                dataset_root + '/exper_20000_noisy_size' + str(current_level_res) + '_mrc').data
            clean_data = mrcfile.open(dataset_root + '/exper_20000_clean_size' + str(current_level_res) + '_mrc').data
            # noisy_data = np.reshape(noisy_data, (25000, 1, current_level_res, current_level_res))
            # clean_data = np.reshape(clean_data, (25000, 1, current_level_res, current_level_res))
            # d_datasets = collect_image_data(dataset_root, minibatch, current_level_res, 2)
            # fake_seed = torch.randn(minibatch, self.latent_size_, 1, 1).cuda()

            # The first step should be keep "stable" status
            # The rest steps would run as two phases, "fadein" -> "stable"
            epoch_start_time = time.time()
            if cursor_pyramid == pyramid_from - 1:
                net_status = "stable"
                for cursor_step in range(stable_steps):
                    self._train(net_level, net_status, net_alpha, minibatch, cursor_step, noisy_data, clean_data)
            else:
                net_status = "fadein"
                for cursor_step in range(fadein_steps):
                    net_alpha = 1.0 - (cursor_step + 1) / fadein_steps
                    self._train(net_level, net_status, net_alpha, minibatch, cursor_step, noisy_data, clean_data)
                for cursor_step in range(stable_steps):
                    net_alpha = 1.0
                    self._train(net_level, "stable", net_alpha, minibatch, cursor_step, noisy_data, clean_data)
            epoch_end_time = time.time()
            print(epoch_end_time - epoch_start_time)
            print()
            torch.save(self.g_net.state_dict(),
                       save_path_name + '/Gnet_%dx%d.pth' % (current_level_res, current_level_res))
            torch.save(self.d_net.state_dict(),
                       save_path_name + '/Dnet_%dx%d.pth' % (current_level_res, current_level_res))
            torch.save(self.e_net.state_dict(),
                       save_path_name + '/Enet_%dx%d.pth' % (current_level_res, current_level_res))
    def _train(self, net_level, net_status, net_alpha, batch_size, cur_step, noisy_data, clean_data):
        # prepare data
        current_level_res = 2 ** (net_level + 2)
        # g_data = fake_seed
        choose_list = [random.randint(0, 19499) for _ in range(batch_size)]
        work_real = clean_data[choose_list]
        work_noisy = noisy_data[choose_list]
        work_real = np.reshape(work_real, (batch_size, 1, current_level_res, current_level_res))
        work_noisy = np.reshape(work_noisy, (batch_size, 1, current_level_res, current_level_res))
        work_noisy = torch.from_numpy(work_noisy).cuda()
        work_real = torch.from_numpy(work_real).cuda()

        # init nets
        self.g_net.net_config = [net_level, net_status, net_alpha]
        self.d_net.net_config = [net_level, net_status, net_alpha]
        self.e_net.net_config = [net_level, net_status, net_alpha]
        inter = self.e_net(work_noisy)
        work_fake = self.g_net(inter)
        d_real = self.d_net(work_real)
        d_fake = self.d_net(work_fake.detach())

        if self.criterion_type_ == "WGAN-GP":
            # update d_net
            self.d_net.zero_grad()
            # # for i in range(CRITIC_FOR_WGANGP):
            mean_real = d_real.mean()
            mean_real.backward(self.gradient_weight_real_)
            mean_fake = d_fake.mean()
            mean_fake.backward(self.gradient_weight_fake_, retain_graph=True)

            gradient_penalty = self._gradient_penalty(work_real.data, work_fake.data, batch_size, current_level_res)
            gradient_penalty.backward()
            self.d_optim.step()

            # update g_net e_net
            self.g_net.zero_grad()
            self.e_net.zero_grad()
            d_fake = self.d_net(work_fake)
            mean_fake = d_fake.mean()
            mean_fake.backward(self.gradient_weight_real_, retain_graph=True)
            l1_loss = opt.lamda * torch.mean(torch.abs(work_fake - work_real))
            l1_loss.backward()
            self.g_optim.step()
            self.e_optim.step()

        elif self.criterion_type_ == "GAN":
            target_real = torch.full((batch_size, 1, 1, 1), 1).cuda()
            target_fake = torch.full((batch_size, 1, 1, 1), 0).cuda()
            # update d_net
            self.d_net.zero_grad()
            loss_d_real = self.criterion(d_real, target_real)
            loss_d_real.backward()
            loss_d_fake = self.criterion(d_fake, target_fake)
            loss_d_fake.backward()
            #loss_d = loss_d_real + loss_d_fake * 0.1
            #loss_d.backward()
            self.d_optim.step()

            # update g_net
            self.g_net.zero_grad()
            d_fake = self.d_net(work_fake)
            loss_g = self.criterion(d_fake, target_real)
            loss_g.backward()
            self.g_optim.step()
            
  

        if cur_step % SAVE_IMG_STEP == 0:
            vutils.save_image(work_real,
                              save_path_name + '/real_netlevel%02d.jpg' % net_level,
                              nrow=4, normalize=True, padding=0)
            fake_work = self.g_net(self.e_net(work_noisy))
            vutils.save_image(fake_work.detach(),
                              save_path_name + '/fake_%s_netlevel%02d_%07d.jpg' % (net_status, net_level, cur_step),
                              nrow=4, normalize=True, padding=0)
            print(torch.mean((work_real - fake_work) ** 2))


    def _init_network(self):
        # Init Generator and Discriminator
        self.g_net = Generator(self.resolution_, self.latent_size_, self.rgb_channel_,
                               self.fmap_base_, self.fmap_decay_, self.fmap_max_, is_tanh=self.is_tanh_)
        self.d_net = Discriminator(self.resolution_, self.rgb_channel_,
                                   self.fmap_base_, self.fmap_decay_, self.fmap_max_, is_sigmoid=self.is_sigmoid_)
        self.e_net = Encoder(self.resolution_, self.rgb_channel_,
                             self.fmap_base_, self.fmap_decay_, self.fmap_max_, is_sigmoid=self.is_sigmoid_)
        # if TO_GPU:
        self.g_net.cuda()
        self.d_net.cuda()
        self.e_net.cuda()

    def _init_optimizer(self):
        self.g_optim = optim.Adam(self.g_net.parameters(), lr=G_LR, betas=ADAM_BETA, eps=ADAM_EPS)
        self.d_optim = optim.Adam(self.d_net.parameters(), lr=D_LR, betas=ADAM_BETA, eps=ADAM_EPS)
        self.e_optim = optim.Adam(self.e_net.parameters(), lr=D_LR, betas=ADAM_BETA, eps=ADAM_EPS)

    def _init_criterion(self):
        if self.criterion_type_ == "GAN":
            self.criterion = nn.BCELoss()
        elif self.criterion_type_ == "WGAN-GP":
            self.criterion = self._gradient_penalty
        else:
            raise ValueError("INVALID TYPE: {0}.".format(self.criterion_type_))

    def _gradient_penalty(self, real_data, fake_data, batch_size, res):
        """
        This algorithm was mentioned on the Page4 of paper
        'Improved Training of Wasserstein GANs'
        This implementation was from 'https://github.com/caogang/wgan-gp'
        """
        epsilon = torch.rand(batch_size, 1)
        epsilon = epsilon.expand(batch_size, int(real_data.nelement() / batch_size)).contiguous().view(batch_size, 1,
                                                                                                       res,
                                                                                                       res)
        epsilon = epsilon.cuda()
        median_x = epsilon * real_data + ((1 - epsilon) * fake_data)

        # if TO_GPU:
        median_x = median_x.cuda()
        median_data = torch.autograd.Variable(median_x, requires_grad=True)

        d_median_data = self.d_net(median_data)

        gradients = torch.autograd.grad(outputs=d_median_data, inputs=median_data,
                                        grad_outputs=torch.ones(d_median_data.size()).cuda(),
                                        create_graph=True, retain_graph=True, only_inputs=True)[0]
        gradients = gradients.view(gradients.size(0), -1)

        gradient_penalty = ((gradients.norm(2, dim=1) - 1) ** 2).mean() * LAMBDA_FOR_WGANGP
        return gradient_penalty


#################################################################################
# Image loading and fake data generating ########################################
#################################################################################
# Temp used, will be rewritten
def collect_image_data(dir, batch_size, resolution, num_workers, max_samplesize=150):
    path = os.path.join(dir, str(resolution))
    dset = vdatasets.ImageFolder(
        root=path, transform=transforms.Compose([
            transforms.ToTensor(),
            transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5)),
        ]))
    dataloader = udata.DataLoader(dset, batch_size=batch_size,
                                  shuffle=True, num_workers=num_workers, drop_last=True)
    output = []
    for i, j in enumerate(dataloader):
        if i == max_samplesize:
            break
        output.append(j)

    return output


def collect_fake_data(tag_dir):
    pass


#################################################################################
# Func for reorganising images ##################################################
#################################################################################
def gen_classified_images(dir_path, classes=None, centre_crop=True, save_to_local=False):
    """
    dir_path
        |_ img1
        |_ img2
    :return:
    """
    classes = [4, 8, 16, 32, 64, 128, 256, 512, 1024] if classes is None else classes
    image_dataset = {i: [] for i in classes}
    for i in os.listdir(dir_path):
        img_path = os.path.join(dir_path, i)
        if not os.path.isfile(img_path):
            continue
        # im = _image_reader(os.path.join(dir_path, i))
        im = Image.open(os.path.join(dir_path, i))
        im_info = im.info
        if im is None:
            continue
        o_width, o_height = im.size
        if centre_crop is True:
            min_dim = o_width if o_width < o_height else o_height
            im = _image_centre_crop(im, (min_dim, min_dim))

        for class_ in classes:
            im_resized = _resize_image(im, (class_, class_))
            image_dataset[class_].append(im_resized)
            if save_to_local is True:
                folder = os.path.join(os.path.dirname(dir_path),
                                      "{0}_{1}".format(os.path.basename(dir_path), "classified"),
                                      str(class_), str(class_))
                os.makedirs(folder, exist_ok=True)
                im_resized.save(os.path.join(folder, i), **im_info)
    return image_dataset


def gen_mrc(dir_path, classes=None, centre_crop=True, save_to_local=False):
    classes = [4, 8, 16, 32, 64, 128, 256] if classes is None else classes
    clean_data = mrcfile.open(os.path.join(dir_path, 'exper_20000_clean.mrc')).data
    noisy_data = mrcfile.open(os.path.join(dir_path, 'exper_20000_noisy.mrc')).data
    for j in classes:
        with mrcfile.new(
                dir_path + '/exper_20000_noisy_size' + str(j) + '_mrc', overwrite=True) as mrc:
            with mrcfile.new(
                    dir_path + '/exper_20000_clean_size' + str(j) + '_mrc', overwrite=True) as mrc1:
                mrc1.set_data(np.zeros((20000, j, j), dtype=np.float32))
                mrc.set_data(np.zeros((20000, j, j), dtype=np.float32))
                for i in range(20000):
                    im_noisy = noisy_data[i]
                    im_clean = clean_data[i]
                    im_noisy = Image.fromarray(im_noisy)
                    im_clean = Image.fromarray(im_clean)
                    mrc.data[i, :, :] = np.array(_resize_image(im_noisy, (j, j)))
                    mrc1.data[i, :, :] = np.array(_resize_image(im_clean, (j, j)))
        print('ok')


def _resize_image(im, size, resample=Image.ANTIALIAS):
    return im.resize(size, resample)


def _image_centre_crop(im, outsize):
    im_width, im_height = im.size
    out_width, out_height = outsize
    left = int(round((im_width - out_width) / 2.))
    upper = int(round((im_height - out_height) / 2.))
    right = left + out_width
    lower = upper + out_height
    return im.crop((left, upper, right, lower))


if __name__ == "__main__":
    p = PGGAN(256, 2, "WGAN-GP")
    p.train('./data')
    #gen_mrc('./data/')