P13.torchvision中的数据集使用

13.1Transforms中的类

1.打开pytorch官网

2.找到CIFAR10，这个数据集比较小

3.点击图片上红色的CIFAR10

4.这里的链接就是Pycharm下载到dataset里面的东西

13.2CIFAR10数据集的下载与导入

1.在Pycharm下载，下载到dataset

2.下载成功的成果

13.3打印输出，查看test_set的构成(img,target)和它的classes

点击查看代码

import torchvision
from torch.utils.tensorboard import SummaryWriter

train_set = torchvision.datasets.CIFAR10(root="./dataset",train=True,download=True)
test_set = torchvision.datasets.CIFAR10(root="./dataset",train=False,download=True)

#多行注释：Ctrl+/
print(test_set[0])  #test_set[i]是由img,target构成的  target是类别
print(test_set.classes)
img,target = test_set[0]
print(img)
print(target)
print(test_set.classes[target])

输出结果如下：

点击查看代码

D:\anaconda3\envs\pytorch\python.exe D:/DeepLearning/Learn_torch/P13_dataset_transform.py
Files already downloaded and verified
Files already downloaded and verified
(<PIL.Image.Image image mode=RGB size=32x32 at 0x1F81317A880>, 3)
['airplane', 'automobile', 'bird', 'cat', 'deer', 'dog', 'frog', 'horse', 'ship', 'truck']
<PIL.Image.Image image mode=RGB size=32x32 at 0x1F81317A8E0>
3
cat

进程已结束,退出代码0

13.4CIFAR10数据集的导入(tensorboard)

1.原始PIL类型转换成tensor数据类型

点击查看代码

原始PIL类型转换成tensor数据类型
dataset_transform = torchvision.transforms.Compose(
    [torchvision.transforms.ToTensor()]
)

2.获取CIFAR10的数据集并对其进行转换

点击查看代码

#将以上transforms的totensor应用到CIFAR10的每一张图片
train_set = torchvision.datasets.CIFAR10(root="./dataset",transform=dataset_transform,train=True,download=True)
test_set = torchvision.datasets.CIFAR10(root="./dataset",transform=dataset_transform,train=False,download=True)

3.写入日志文件“p13”,将数据集导入到tensorboard进行显示

点击查看代码

writer = SummaryWriter("P13")
for i in range(10):
    img,target = test_set[i]
    writer.add_image("test_set",img,i)
    #i：global_step（一个整数，通常表示训练的步数或者迭代次数等，用于在记录多张图像时区分不同阶段的图像）
writer.close()

输出结果如下：

点击查看代码

D:\anaconda3\envs\pytorch\python.exe D:/DeepLearning/Learn_torch/P13_dataset_transform.py
Files already downloaded and verified
Files already downloaded and verified

进程已结束,退出代码0

4.在终端中打开

tensorboard --logdir=D:\DeepLearning\Learn_torch\P13

5.点击网址

【注：整体代码如下：】

点击查看代码

import torchvision
from torch.utils.tensorboard import SummaryWriter
#原始PIL类型转换成tensor数据类型
dataset_transform = torchvision.transforms.Compose(
    [torchvision.transforms.ToTensor()]
)

# train_set = torchvision.datasets.CIFAR10(root="./dataset",train=True,download=True)
# test_set = torchvision.datasets.CIFAR10(root="./dataset",train=False,download=True)

#将以上transforms的totensor应用到CIFAR10的每一张图片
train_set = torchvision.datasets.CIFAR10(root="./dataset",transform=dataset_transform,train=True,download=True)
test_set = torchvision.datasets.CIFAR10(root="./dataset",transform=dataset_transform,train=False,download=True)

# #多行注释：Ctrl+/
# print(test_set[0])  #test_set[i]是由img,target构成的  target是类别
# print(test_set.classes)
# img,target = test_set[0]
# print(img)
# print(target)
# print(test_set.classes[target])


writer = SummaryWriter("P13")
for i in range(10):
    img,target = test_set[i]
    writer.add_image("test_set",img,i)
    #i：global_step（一个整数，通常表示训练的步数或者迭代次数等，用于在记录多张图像时区分不同阶段的图像）
writer.close()