image
imagewidth (px) 28
28
| text
stringclasses 20
values |
|---|---|
5
|
|
0
|
|
3
|
|
5
|
|
0
|
|
3
|
|
seven
|
|
seven
|
|
zero
|
|
zero
|
|
one
|
|
three
|
|
nine
|
|
one
|
|
five
|
|
six
|
|
8
|
|
six
|
|
eight
|
|
eight
|
|
four
|
|
five
|
|
one
|
|
eight
|
|
seven
|
|
seven
|
|
six
|
|
7
|
|
three
|
|
five
|
|
eight
|
|
six
|
|
one
|
|
four
|
|
nine
|
|
two
|
|
nine
|
|
seven
|
|
9
|
|
four
|
|
eight
|
|
five
|
|
one
|
|
six
|
|
three
|
|
six
|
|
two
|
|
zero
|
|
zero
|
|
9
|
|
zero
|
|
five
|
|
zero
|
|
zero
|
|
two
|
|
one
|
|
seven
|
|
nine
|
|
eight
|
|
two
|
|
0
|
|
nine
|
|
three
|
|
one
|
|
two
|
|
two
|
|
three
|
|
one
|
|
five
|
|
one
|
|
five
|
|
1
|
|
seven
|
|
seven
|
|
four
|
|
nine
|
|
eight
|
|
nine
|
|
four
|
|
seven
|
|
three
|
|
one
|
|
1
|
|
seven
|
|
one
|
|
six
|
|
nine
|
|
zero
|
|
zero
|
|
two
|
|
seven
|
|
three
|
|
three
|
|
5
|
|
seven
|
|
four
|
|
zero
|
|
eight
|
|
two
|
|
six
|
MNIST for Diffusion
Training a diffusion model from scratch is pretty cool, why not do so with the canonical "hello world" dataset of computer vision? This dataset matches the sample dataset from this text_to_image.py diffusion tutorial. Specifying ckg/mnist-for-diffusion ought get you off to the races.
This dataset contains two copies of the original MNIST train & test sets. The first half of the dataset contains MNIST images with the string-ified class id (i.e: "1") and the second half has the class id mapped to a natural language name (i.e: "one"). This little data augmentation doubles the number of samples and should result in interesting behavior if you train a U-Net from scratch whilst using a frozen, pre-trained text-encoder!
Thank you LeCun & Cortes for making this dataset available.
- Downloads last month
- 9