Datasets:

ylecun

/

mnist

like 84

Hi there, is possible if i wanna load MNIS dataset using load_dataset function and got the directory of the image with corresponding labels folder?

for example i load with this

from datasets import load_dataset

dataset = load_dataset("mnist", split="train[:100]")

and return of dataset is like beloow

Dataset({
    features: ['image', 'label'],
    num_rows: 100
})

but, in my case i wanna get the directory of image and coresponding labels folder like this

MNIST_Dataset/
|__ 0/
|    |-- image_1.png
|    |-- image_2.png
|    |__ ...
|
|__ 1/
|    |-- image_3.png
|    |-- image_4.png
|    |__ ...

How to get like that?