parser = argparse.ArgumentParser(description='Train Super Resolution Models')
parser.add_argument('--crop_size', default=88, type=int, help='training images crop size')
parser.add_argument('--upscale_factor', default=4, type=int, choices=[2, 4, 8],
help='super resolution upscale factor')
parser.add_argument('--num_epochs', default=100, type=int, help='train epoch number')
parser.add_argument('--loss_net', default="vgg16", type=string, help='choose network for loss')
if __name__ == '__main__':
opt = parser.parse_args()
CROP_SIZE = opt.crop_size
UPSCALE_FACTOR = opt.upscale_factor
NUM_EPOCHS = opt.num_epochs
train_set = TrainDatasetFromFolder('data/DIV2K_train_HR', crop_size=CROP_SIZE, upscale_factor=UPSCALE_FACTOR)
val_set = ValDatasetFromFolder('data/DIV2K_valid_HR', upscale_factor=UPSCALE_FACTOR)
train_loader = DataLoader(dataset=train_set, num_workers=4, batch_size=64, shuffle=True)
val_loader = DataLoader(dataset=val_set, num_workers=4, batch_size=1, shuffle=False)
netG = Generator(UPSCALE_FACTOR)
print('# generator parameters:', sum(param.numel() for param in netG.parameters()))
netD = Discriminator()
print('# discriminator parameters:', sum(param.numel() for param in netD.parameters()))
generator_criterion = GeneratorLoss(opt.loss_net)
if torch.cuda.is_available():
netG.cuda()
netD.cuda()
generator_criterion.cuda()
help='super resolution upscale factor')
parser.add_argument('--num_epochs',
default=100,
type=int,
help='train epoch number')
opt = parser.parse_args()
CROP_SIZE = opt.crop_size
UPSCALE_FACTOR = opt.upscale_factor
NUM_EPOCHS = opt.num_epochs
train_set = TrainDatasetFromFolder('data/VOC2012/celeba',
crop_size=CROP_SIZE,
upscale_factor=UPSCALE_FACTOR)
val_set = ValDatasetFromFolder('data/VOC2012/val',
upscale_factor=UPSCALE_FACTOR)
train_loader = DataLoader(dataset=train_set,
num_workers=4,
batch_size=400,
shuffle=True)
val_loader = DataLoader(dataset=val_set,
num_workers=4,
batch_size=1,
shuffle=False)
netG = Generator(UPSCALE_FACTOR)
print('# generator parameters:',
sum(param.numel() for param in netG.parameters()))
netD = Discriminator()
print('# discriminator parameters:',
sum(param.numel() for param in netD.parameters()))
def main():
os.environ["CUDA_VISIBLE_DEVICES"] = '0, 1, 2, 3'
train_set = TrainDatasetFromFolder('data/DIV2K_train_HR',
crop_size=CROP_SIZE,
upscale_factor=UPSCALE_FACTOR)
val_set = ValDatasetFromFolder('data/DIV2K_valid_HR',
upscale_factor=UPSCALE_FACTOR)
train_loader = DataLoader(dataset=train_set,
num_workers=4,
batch_size=64,
shuffle=True)
val_loader = DataLoader(dataset=val_set,
num_workers=4,
batch_size=1,
shuffle=False)
netG = Generator(UPSCALE_FACTOR)
print('# generator parameters:',
sum(param.numel()
for param in netG.parameters())) # Generator의 총 parameter 수
netD = Discriminator()
print('# discriminator parameters:',
sum(param.numel()
for param in netD.parameters())) # Discriminator의 총 parameter 수
generator_criterion = GeneratorLoss() # loss function
netG = nn.DataParallel(netG).cuda()
netD = nn.DataParallel(netD).cuda()
#netG = netG.cuda()
#netD = netD.cuda()
generator_criterion.cuda()
optimizerG = optim.Adam(netG.parameters()) # optimizer : adam
optimizerD = optim.Adam(netD.parameters()) # optimizer : adam
results = {
'd_loss': [],
'g_loss': [],
'd_score': [],
'g_score': [],
'psnr': [],
'ssim': []
}
for epoch in range(1, NUM_EPOCHS + 1):
# model train
d_loss, g_loss, d_score, g_score = train(netG, netD,
generator_criterion,
optimizerG, optimizerD,
train_loader, epoch)
# validation data acc
psnr, ssim = test(netG, netD, val_loader, epoch)
# save loss\scores\psnr\ssim
results['d_loss'].append(d_loss)
results['g_loss'].append(g_loss)
results['d_score'].append(d_score)
results['g_score'].append(g_score)
results['psnr'].append(psnr)
results['ssim'].append(ssim)
# save results
if epoch % 10 == 0 and epoch != 0:
out_path = 'statistics/'
data_frame = pd.DataFrame(data={
'Loss_D': results['d_loss'],
'Loss_G': results['g_loss'],
'Score_D': results['d_score'],
'Score_G': results['g_score'],
'PSNR': results['psnr'],
'SSIM': results['ssim']
},
index=range(1, epoch + 1))
data_frame.to_csv(out_path + 'srf_' + str(UPSCALE_FACTOR) +
'_train_results.csv',
index_label='Epoch')
print()
type=int,
help='train epoch number')
# 对之前add的参数进行赋值，并返回响应namespace
opt = parser.parse_args()
# 提取opt（选项器）中设置的参数，设定为常量
CROP_SIZE = opt.crop_size
UPSCALE_FACTOR = opt.upscale_factor
NUM_EPOCHS = opt.num_epochs
# 从指定路径导入train_set，指定裁剪大小和放大因子
train_set = TrainDatasetFromFolder('./data/train',
crop_size=CROP_SIZE,
upscale_factor=UPSCALE_FACTOR)
val_set = ValDatasetFromFolder('./data/valLR', upscale_factor=UPSCALE_FACTOR)
# 使用loader，从训练集中，一次性处理一个batch的文件 （批量加载器）
train_loader = DataLoader(dataset=train_set,
num_workers=8,
batch_size=64,
shuffle=True)
val_loader = DataLoader(dataset=val_set,
num_workers=8,
batch_size=1,
shuffle=False)
# 创建生成器实例 netG ，输出生成器参数的数量
netG = Generator(UPSCALE_FACTOR)
netG = nn.DataParallel(netG) #,device_ids=[0,1,2]) # multi-GPU
print('# generator parameters:',
sum(param.numel() for param in netG.parameters()))
default=100,
type=int,
help='train epoch number')
if __name__ == '__main__':
opt = parser.parse_args()
CROP_SIZE = opt.crop_size
UPSCALE_FACTOR = opt.upscale_factor
NUM_EPOCHS = opt.num_epochs
train_set = TrainDatasetFromFolder('data/DIV2K_train_HR',
crop_size=CROP_SIZE,
upscale_factor=UPSCALE_FACTOR)
val_set = ValDatasetFromFolder('data/DIV2K_valid_HR',
crop_size=CROP_SIZE * 2,
upscale_factor=UPSCALE_FACTOR)
train_loader = DataLoader(dataset=train_set,
num_workers=4,
batch_size=4,
shuffle=True)
val_loader = DataLoader(dataset=val_set,
num_workers=4,
batch_size=1,
shuffle=False)
netG = Generator(UPSCALE_FACTOR)
print('# generator parameters:',
sum(param.numel() for param in netG.parameters()))
netD = Discriminator()
print('# discriminator parameters:',