default_loader 在 torch 中的作用是什么？

Question

    import os
    import pandas as pd
    import numpy as np
    from torchvision.datasets.folder import default_loader
    from torchvision.datasets.utils import download_url
    from torch.utils.data import Dataset

   
 class Sample_Class(Dataset):
      def __init__(self,root,train=True,transform=None,loader=default_loader):
        self.root = os.path.expanduser(root)
        self.transform = transform
        self.loader = default_loader

上面的代码片段中，loader=default_loader的意义是什么，具体是做什么的？

Answer 1

这个 Sample_Class 很可能是在模仿 ImageFolder, DatasetFolder, and ImageNet 的行为。该函数应将文件名作为输入，return 是 PIL.Image 或 accimage.Image，具体取决于所选的图像后端。

default_loader函数定义在torchvision/datasets/folder.py

def pil_loader(path):
    # open path as file to avoid ResourceWarning (https://github.com/python-pillow/Pillow/issues/835)
    with open(path, 'rb') as f:
        img = Image.open(f)
        return img.convert('RGB')


def accimage_loader(path):
    import accimage
    try:
        return accimage.Image(path)
    except IOError:
        # Potentially a decoding problem, fall back to PIL.Image
        return pil_loader(path)


def default_loader(path):
    from torchvision import get_image_backend
    if get_image_backend() == 'accimage':
        return accimage_loader(path)
    else:
        return pil_loader(path)

注意：default_loader 默认为 PIL reader

default_loader 在 torch 中的作用是什么？

what is the function of default_loader in torch?

dataset

pytorch

dataloader