图像分类数据集 (热狗 hotdog)

引入

  热狗数据集包含2800包含热狗和其他食物的图像，下载地址如下：
  原始：https://apache-mxnet.s3-accelerate.amazonaws.com/gluon/dataset/hotdog.zip
  微云：链接：https://share.weiyun.com/lTrk8xq4 密码：fks7u4

1 图像载入

from torchvision.datasets import ImageFolder


def load_image(save_home="D:/Data/Image/hotdog/"):
    """
    图像载入
    """
    # 训练集和测试集
    return ImageFolder(save_home + "train"), ImageFolder(save_home + "test")


def test1():
    """测试之图像载入"""
    image_tr, image_te = load_image()


if __name__ == '__main__':
    test1()

2 图像展示

def test2():
    """测试之图像展示"""

    def show_images(imgs, num_rows, num_cols, scale=1.4):
        import matplotlib.pyplot as plt
        _, axes = plt.subplots(num_rows, num_cols, figsize=(num_cols * scale, num_rows * scale))
        for i in range(num_rows):
            for j in range(num_cols):
                axes[i][j].imshow(imgs[i * num_cols + j])
                axes[i][j].axes.get_xaxis().set_visible(False)
                axes[i][j].axes.get_yaxis().set_visible(False)
        plt.show()

    image_tr, image_te = load_image()
    hotdogs = [image_tr[i][0] for i in range(8)]
    not_hotdogs = [image_tr[-i - 1][0] for i in range(8)]
    show_images(hotdogs + not_hotdogs, 2, 8)


if __name__ == '__main__':
    test2()

  输出如下：

致谢

  感谢李沐、Aston Zhang等老师的这本《动手学深度学习》一书，为鄙人学习深度学习提供了很大的帮助。本文一系列关于深度学习的博客均无侵权之意，只为记录自己的深度学习历程。
  项目Github地址：https://github.com/ShusenTang/Dive-into-DL-PyTorch