pytorch 数据集预处理

pytorch 数据集预处理

1、定义transforms

train_tfm = transforms.Compose([
    # transforms.ToPILImage(),
    # Resize the image into a fixed shape (height = width = 128)
    # transforms.Resize((128, 128)),
    # You may add some transforms here.
    # ToTensor() should be the last one of the transforms.
    transforms.ToTensor(),  #具有维度转化功能
    transforms.Normalize(mean=[57.53], std=[35.825])
])

2、定义train_loader、test_loader

totall_set = DatasetFolder("D:/PythonProject/QTNLS/resources/TN-Face/labeled", loader=tifffile.imread, extensions="tif", transform=train_tfm)
# totall_set = DatasetFolder("D:/PythonProject/QTNLS/resources/TN-Face/labeled", loader=tifffile.imread, extensions="tif", transform=None)
size_train = int(0.6 * len(totall_set))
size_test = len(totall_set) - size_train

train_set, test_set = torch.utils.data.random_split(totall_set, [size_train, size_test])

train_loader = DataLoader(train_set, batch_size=batch_size_train, shuffle=True)
test_loader = DataLoader(test_set, batch_size=batch_size_test, shuffle=True)

3、在训练的时候，还需要进行数据维度操纵

data = data.permute(0,2,1,3) #对维度重新进行调整
data = data.unsqueeze(1)

4、同理，测试的test的时候，也需要。