num_gpu: 使用的 GPU 数量。 manual_seed: 随机种子。 数据集和数据加载器设置 (Dataset and Data Loader Settings): datasets: 包括训练和验证数据集的设置。 gt_size: 地面真值的大小。 use_hflip, use_rot: 数据增强选项。 batch_size_per_gpu: 每个 GPU 的批量大小。 网络结构 (Network Structures): ...
name: IMDN_x4_test_DIV2K model_type: SRModel scale: 4 num_gpu: 1 # set num_gpu: 0 for cpu mode manual_seed: 0 datasets: test_1: name: DIV2K100 type: PairedImageDataset dataroot_gt: D:/Datasets/SISR/DIV2K/DIV2K_valid_HR dataroot_lq: D:/Datasets/SISR/DIV2K/DIV2K_valid_LR_...
num_gpu: 1 # set num_gpu: 0 for cpu mode manual_seed: 0 # dataset and data loader settings datasets: train: name: face_dataset type: PairedImageDataset dataroot_gt: data/hq dataroot_lq: data/lq filename_tmpl: '{}' io_backend: type: disk gt_size: 384 use_flip: true use_rot:...
true use_rot: true # --- arguments of data loader --- # use_shuffle: true num_worker_per_gpu: 3 batch_size_per_gpu: 16 dataset_enlarge_ratio: 10 prefetch_mode: ~ val: # validation dataset name: ExampleSet5 type: ExampleDataset dataroot_gt: datasets/example/Set5 io_backend: type:...
# 实验名称 name: 001_MSRResNet_x4_f64b16_DIV2K_1000k_B16G1_wandb # 使用的model类型, 一般为在`models`目录下定义的模型的类名 model_type: SRModel # 输出相比输入的放大比率, 在SR中是放大倍数; 若有些任务没有这个配置, 则写1 scale: 4 # 测试卡数 num_gpu: 1 # set num_gpu: 0 for ...
它会多占用一些GPU显存. 注意: 这个模式下, 一定要设置 pin_memory=True prefetch_mode: cuda pin_memory: true prefetch_mode: cpu. 使用 CPU prefetcher, 具体介绍参见 IgorSusmelj/pytorch-styleguide. (目前测试,这个加速不明显) prefetch_mode: cpu num_prefetch_queue: 1 # 1 by default 图像数据 ...
diskgt_size:128use_flip:trueuse_rot:true# --- arguments of data loader --- #use_shuffle:truenum_worker_per_gpu:3batch_size_per_gpu:16dataset_enlarge_ratio:10prefetch_mode:~val:# validation datasetname:ExampleSet5type:ExampleDatasetdataroot_gt:datasets/example/Set5io_backend:type:disk ②...
num_gpu=opt['num_gpu'], dist=opt['dist'], sampler=train_sampler, seed=opt['manual_seed']) num_iter_per_epoch = math.ceil( len(train_set) * dataset_enlarge_ratio / (dataset_opt['batch_size_per_gpu'] * opt['world_size'])) ...
Open Source Image and Video Restoration Toolbox for Super-resolution, Denoise, Deblurring, etc. Currently, it includes EDSR, RCAN, SRResNet, SRGAN, ESRGAN, EDVR, BasicVSR, SwinIR, ECBSR, etc. Also support StyleGAN2, DFDNet. - BasicSR/basicsr/models/video
device('cuda' if opt['num_gpu'] != 0 else 'cpu') self.is_train = opt['is_train'] self.schedulers = [] self.optimizers = [] def feed_data(self, data): pass def optimize_parameters(self): pass def get_current_visuals(self): pass def save(self, epoch, current_...