Update README.md
Browse files
README.md
CHANGED
@@ -7,13 +7,25 @@ dataset_info:
|
|
7 |
dtype: string
|
8 |
splits:
|
9 |
- name: train
|
10 |
-
num_bytes: 39858266
|
11 |
num_examples: 140000
|
12 |
download_size: 37136812
|
13 |
-
dataset_size: 39858266
|
14 |
configs:
|
15 |
- config_name: default
|
16 |
data_files:
|
17 |
- split: train
|
18 |
path: data/train-*
|
|
|
|
|
|
|
19 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
7 |
dtype: string
|
8 |
splits:
|
9 |
- name: train
|
10 |
+
num_bytes: 39858266
|
11 |
num_examples: 140000
|
12 |
download_size: 37136812
|
13 |
+
dataset_size: 39858266
|
14 |
configs:
|
15 |
- config_name: default
|
16 |
data_files:
|
17 |
- split: train
|
18 |
path: data/train-*
|
19 |
+
license: cc
|
20 |
+
size_categories:
|
21 |
+
- 100K<n<1M
|
22 |
---
|
23 |
+
|
24 |
+
|
25 |
+
# MNIST for Diffusion
|
26 |
+
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](https://github.com/huggingface/diffusers/tree/main/examples/text_to_image). Specifying `ckg/mnist-for-diffusion` ought get you off to the races.
|
27 |
+
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!
|
28 |
+
|
29 |
+
|
30 |
+
|
31 |
+
Thank you LeCun & Cortes for making this dataset available.
|