from torchvision import transforms import torchvision.utils as vutils from tensorboardX import SummaryWriter from data.data import InpaintingDataset, ToTensor from model.net import InpaintingModel_DFBM from options.train_options import TrainOptions from util.utils import getLatest from multiprocessing import freeze_support if __name__ == '__main__': config = TrainOptions().parse() print('loading data..') dataset = InpaintingDataset(config.data_file, config.dataset_path, transform=transforms.Compose([ToTensor()])) dataloader = DataLoader(dataset, batch_size=config.batch_size, shuffle=True, num_workers=4, drop_last=True) print('data loaded..') print('configuring model..') ourModel = InpaintingModel_DFBM(opt=config) ourModel.print_networks() if config.load_model_dir != '': print('Loading pretrained model from {}'.format(config.load_model_dir)) ourModel.load_networks( getLatest(os.path.join(config.load_model_dir, '*.pth')))
py_file_name = config.py_file.split("/")[ -1] # Get python file name (soruce code name) checkpoint_dir = os.path.join(config.out_dir, py_file_name + "/checkpoints") os.makedirs(checkpoint_dir, exist_ok=True) # make tensorboard subdirectory for the experiment tensorboard_exp_dir = os.path.join(config.tensorboard_dir, py_file_name) os.makedirs(tensorboard_exp_dir, exist_ok=True) #======================================== print('loading data..') dataset = InpaintingDataset_WithMask_v2(config.dataset_path, config.data_root, transform=transforms.Compose( [ToTensor()])) dataloader = DataLoader(dataset, batch_size=config.batch_size, shuffle=True, num_workers=4, drop_last=True) print('data loaded..') print('configuring model..') ourModel = InpaintingModel_GMCNN_Given_Mask(in_channels=4, opt=config) ourModel.print_networks() if (config.load_model_dir != '') or (config.load_model_path != ''): print('Loading pretrained model from {}'.format(config.load_model_dir)) #ourModel.load_networks(getLatest(os.path.join(config.load_model_dir, '*.pth'))) # Modified by Vajira to load exact path ourModel.load_networks(config.load_model_path)