GoodWin/Deep-Multi-scale

1from .base_model import BaseModel2from . import networks3import torch4import numpy as np5import torchvision.transforms as transforms6import PIL7 8import torch.nn.functional as F9 10class TestModel(BaseModel):11    def name(self):12        return 'TestModel'13 14    @staticmethod15    def modify_commandline_options(parser, is_train=True):16        assert not is_train, 'TestModel cannot be used in train mode'17        parser.set_defaults(dataset_mode='aligned')18 19        parser.add_argument('--model_suffix', type=str, default='',20                            help='In checkpoints_dir, [which_epoch]_net_G[model_suffix].pth will'21                            ' be loaded as the generator of TestModel')22        return parser23 24    def initialize(self, opt):25        assert(not opt.isTrain)26        BaseModel.initialize(self, opt)27 28        # specify the training losses you want to print out. The program will call base_model.get_current_losses29        self.loss_names = []30        # specify the images you want to save/display. The program will call base_model.get_current_visuals31        self.visual_names = ['fake_A','real_A']32        self.model_names = ['G']33 34        self.netG = networks.define_G('UNetDictFace',self.gpu_ids)35 36    def set_input(self, input):37        self.real_A = input['A'].to(self.device) #degraded img38        self.real_C = input['C'].to(self.device) #groundtruth39        self.image_paths = input['A_paths']40        self.Part_locations = input['Part_locations']41 42    def forward(self):43        44        self.fake_A = self.netG(self.real_A, self.Part_locations) #45        # try:46        #     self.fake_A = self.netG(self.real_A, self.Part_locations) #生成图47        # except:48        #     self.fake_A = self.real_A49