Pytorch专题实战——数据转换(Dataset Transforms)

1.导入必要模块

import torch
from torch.utils.data import Dataset
import numpy as np
import torchvision

2.定义数据处理类

class WineDataset(Dataset):
    def __init__(self, transform=None):
        xy = np.loadtxt('./wine.csv',delimiter=',',dtype=np.float32,skiprows=1)
        self.n_samples = xy.shape[0]
        
        self.x_data = xy[:,1:]
        self.y_data = xy[:,[0]]
        
        self.transform = transform
    
    def __getitem__(self, index):
        sample = self.x_data[index], self.y_data[index]
        
        if self.transform:
            sample = self.transform(sample)
        return sample
    
    def __len__(self):
        return self.n_samples

3.定义numpy转化为tensor类

class ToTensor:
    def __call__(self, sample):   #可调用对象
        inputs, targets = sample
        return torch.from_numpy(inputs), torch.from_numpy(targets)

4.定义乘法转化类

class MulTransform:
    def __init__(self, factor):
        self.factor = factor
    
    def __call__(self, sample):
        inputs, targets = sample
        inputs *= self.factor    #数据*影响因子
        return inputs, targets

5.打印结果

5.1.未转化前

print("Without Transform")
dataset = WineDataset()
first_data = dataset[0]
features, labels = first_data
print(type(features), type(labels))
print(features, labels)

5.2.tensor转化

print("with Tensor Transform")
dataset = WineDataset(transform=ToTensor())
first_data = dataset[0]
features, labels = first_data
print(type(features), type(labels))
print(features, labels)

5.3.乘法转化

print('with Tensor and Multiplication Transform')
composed = torchvision.transforms.Compose([ToTensor(), MulTransform(4)])
dataset = WineDataset(transform=composed)
first_data = dataset[0]
features, labels = first_data
print(type(features), type(labels))
print(features, labels)