def init_models(opt, img_shape) -> "Tuple[models.WDiscriminator, models.WDiscriminator, models.WDiscriminator, models.GeneratorConcatSkip2CleanAdd]":
#generator initialization:
netG = models.GeneratorConcatSkip2CleanAdd(opt, img_shape).to(opt.device)
netG.apply(models.weights_init)
if opt.netG != '':
netG.load_state_dict(torch.load(opt.netG))
logger.info(netG)
# general discriminator initialization for both images:
netD = models.WDiscriminator(opt).to(opt.device)
netD.apply(models.weights_init)
if opt.netD != '':
netD.load_state_dict(torch.load(opt.netD))
logger.info(netD)
# discriminator initialization for identifying the mask of the first image:
netD_mask1 = models.WDiscriminator(opt).to(opt.device)
netD_mask1.apply(models.weights_init)
if opt.netD_mask1 != '':
netD_mask1.load_state_dict(torch.load(opt.netD_mask1))
logger.info(netD_mask1)
# discriminator initialization for identifying the mask of the second image:
netD_mask2 = models.WDiscriminator(opt).to(opt.device)
netD_mask2.apply(models.weights_init)
if opt.netD_mask2 != '':
netD_mask2.load_state_dict(torch.load(opt.netD_mask2))
logger.info(netD_mask2)
return netD, netD_mask1, netD_mask2, netG
def init_models(opt):
# 模型初始化
netG = models.GeneratorConcatSkip2CleanAdd(opt).to(opt.device)
netG.apply(models.weights_init)
if opt.netG != '':
netG.load_state_dict(torch.load(opt.netG))
print(netG)
# discriminator initialization:
netD = models.WDiscriminator(opt).to(opt.device)
netD.apply(models.weights_init)
if opt.netD != '':
netD.load_state_dict(torch.load(opt.netD))
print(netD)
return netD, netG
def init_models(opt):
# opt.min_nfc is the main parameter that controls the width (filter number) for each layer
#generator initialization:
netG = models.GeneratorConcatSkip2CleanAdd(opt).to(opt.device)
netG.apply(
models.weights_init) # apply weight initialize function of models.
if opt.netG != '':
netG.load_state_dict(torch.load(opt.netG))
print(netG)
#discriminator initialization:
netD = models.WDiscriminator(opt).to(opt.device)
netD.apply(models.weights_init)
if opt.netD != '':
netD.load_state_dict(torch.load(opt.netD))
print(netD)
return netD, netG
def init_models(opt):
netD = models.WDiscriminator(opt)
netG = models.GeneratorConcatSkip2CleanAdd(opt)
return netD, netG
