小土堆pytorch教程学习笔记P10P11

P10.Transforms的使用（一）torchvision中的transforms

1.工程文件夹下存储一些图片

2.tansforms结构及用法

transforms.py工具箱（totensor，resize...)

将一些特定格式的图片经过工具输出结果

通过transform.ToTensor去看两个问题（transforms该如何使用，为什么我们需要Tensor数据类型)

class ToTensor(object):
    """Convert a ``PIL Image`` or ``numpy.ndarray`` to tensor.

def __call__(self, pic):
    """
    Args:
        pic (PIL Image or numpy.ndarray): Image to be converted to tensor.

    Returns:
        Tensor: Converted image.
    """
    return F.to_tensor(pic)

from PIL import Image
from torchvision import transforms

img_path = "dataset/train/ants_image/0013035.jpg"
img = Image.open(img_path)
print(img)

#运行结果：图片被正确读取
 

transforms的使用

from PIL import Image
from torchvision import transforms

img_path = "dataset/train/ants_image/0013035.jpg"
img = Image.open(img_path)

tensor_trans = transforms.ToTensor()
tensor_img = tensor_trans(img)

print(tensor_img)

tensor([[[0.3137, 0.3137, 0.3137,  ..., 0.3176, 0.3098, 0.2980],
         [0.3176, 0.3176, 0.3176,  ..., 0.3176, 0.3098, 0.2980],
         [0.3216, 0.3216, 0.3216,  ..., 0.3137, 0.3098, 0.3020],
         ...,
         [0.3412, 0.3412, 0.3373,  ..., 0.1725, 0.3725, 0.3529],
         [0.3412, 0.3412, 0.3373,  ..., 0.3294, 0.3529, 0.3294],
         [0.3412, 0.3412, 0.3373,  ..., 0.3098, 0.3059, 0.3294]],

        [[0.5922, 0.5922, 0.5922,  ..., 0.5961, 0.5882, 0.5765],
         [0.5961, 0.5961, 0.5961,  ..., 0.5961, 0.5882, 0.5765],
         [0.6000, 0.6000, 0.6000,  ..., 0.5922, 0.5882, 0.5804],
         ...,
         [0.6275, 0.6275, 0.6235,  ..., 0.3608, 0.6196, 0.6157],
         [0.6275, 0.6275, 0.6235,  ..., 0.5765, 0.6275, 0.5961],
         [0.6275, 0.6275, 0.6235,  ..., 0.6275, 0.6235, 0.6314]],

        [[0.9137, 0.9137, 0.9137,  ..., 0.9176, 0.9098, 0.8980],
         [0.9176, 0.9176, 0.9176,  ..., 0.9176, 0.9098, 0.8980],
         [0.9216, 0.9216, 0.9216,  ..., 0.9137, 0.9098, 0.9020],
         ...,
         [0.9294, 0.9294, 0.9255,  ..., 0.5529, 0.9216, 0.8941],
         [0.9294, 0.9294, 0.9255,  ..., 0.8863, 1.0000, 0.9137],
         [0.9294, 0.9294, 0.9255,  ..., 0.9490, 0.9804, 0.9137]]])

 P11.Transforms的使用（二）

pip install opencv-python #打开terminal安装opencv
Requirement already satisfied: opencv-python in /home/ydd/anaconda3/envs/pytorch/lib/python3.6/site-packages (4.5.5.64)
Requirement already satisfied: numpy>=1.13.3 in /home/ydd/anaconda3/envs/pytorch/lib/python3.6/site-packages (from opencv-python) (1.19.5)

#转换为numpy数据类型
import cv2
cv_img = cv2.imread(img_path)

from PIL import Image
from torch.utils.tensorboard import SummaryWriter
from torchvision import transforms

img_path = "dataset/train/ants_image/0013035.jpg"
img = Image.open(img_path)

writer = SummaryWriter("logs")

tensor_trans = transforms.ToTensor()
tensor_img = tensor_trans(img)

writer.add_image("Tensor_img", tensor_img)

writer.close()

 tensorboard --logdir=logs #在terminal中输入，打开事件文件