diffusiondb

Name: diffusiondb
Author: poloclub

1.4kstars79forks

How to install diffusiondb?

poloclub/diffusiondb

$git clone https://github.com/poloclub/diffusiondb

Installs into the current project.

Summary

DiffusionDB is the first large-scale text-to-image prompt dataset. It contains 14 million images generated by Stable Diffusion using prompts and hyperparameters specified by real users. The unprecedented scale and diversity of this human-actuated dataset provide exciting rese

Files · 1

View on GitHub

README.md

1	# DiffusionDB <a href="https://huggingface.co/datasets/poloclub/diffusiondb"><picture><source media="(prefers-color-scheme: dark)" srcset="https://i.imgur.com/yGxUUlX.png"><img src="favicon.ico" align="right" src="favicon.ico" height="40px"></picture>
2
3	[![hugging](https://img.shields.io/badge/🤗%20Hugging%20Face-Datasets-yellow)](https://huggingface.co/datasets/poloclub/diffusiondb)
4	[![license](https://img.shields.io/badge/License-CC0/MIT-blue)](#licensing)
5	[![arxiv badge](https://img.shields.io/badge/arXiv-2210.14896-red)](https://arxiv.org/abs/2210.14896)
6	[![datasheet](https://img.shields.io/badge/Data%20Sheet-Available-success)](https://poloclub.github.io/diffusiondb/datasheet.html)
7	<!-- [![DOI:10.1145/3491101.3519653](https://img.shields.io/badge/DOI-10.1145/3491101.3519653-blue)](https://doi.org/10.1145/3491101.3519653) -->
8
9	<img width="100%" src="https://user-images.githubusercontent.com/15007159/201762588-f24db2b8-dbb2-4a94-947b-7de393fc3d33.gif">
10
11	DiffusionDB is the first large-scale text-to-image prompt dataset. It contains 14 million images generated by Stable Diffusion using prompts and hyperparameters specified by real users. The unprecedented scale and diversity of this human-actuated dataset provide exciting research opportunities in understanding the interplay between prompts and generative models, detecting deepfakes, and designing human-AI interaction tools to help users more easily use these models.
12
13	## Get Started
14
15	DiffusionDB is available at [🤗 Hugging Face Datasets](https://huggingface.co/datasets/poloclub/diffusiondb).
16
17	## Two Subsets
18
19	DiffusionDB provides two subsets (DiffusionDB 2M and DiffusionDB Large) to support different needs.
20
21	\|Subset\|Num of Images\|Num of Unique Prompts\|Size\|Image Directory\|Metadata Table\|
22	\|:--\|--:\|--:\|--:\|--:\|--:\|
23	\|DiffusionDB 2M\|2M\|1.5M\|1.6TB\|`images/`\|`metadata.parquet`\|
24	\|DiffusionDB Large\|14M\|1.8M\|6.5TB\|`diffusiondb-large-part-1/` `diffusiondb-large-part-2/`\|`metadata-large.parquet`\|
25
26	##### Key Differences
27
28	1. Two subsets have a similar number of unique prompts, but DiffusionDB Large has much more images. DiffusionDB Large is a superset of DiffusionDB 2M.
29	2. Images in DiffusionDB 2M are stored in `png` format; images in DiffusionDB Large use a lossless `webp` format.
30
31	## Dataset Structure
32
33	We use a modularized file structure to distribute DiffusionDB. The 2 million images in DiffusionDB 2M are split into 2,000 folders, where each folder contains 1,000 images and a JSON file that links these 1,000 images to their prompts and hyperparameters. Similarly, the 14 million images in DiffusionDB Large are split into 14,000 folders.
34
35	```bash
36	# DiffusionDB 2M
37	./
38	├── images
39	│ ├── part-000001
40	│ │ ├── 3bfcd9cf-26ea-4303-bbe1-b095853f5360.png
41	│ │ ├── 5f47c66c-51d4-4f2c-a872-a68518f44adb.png
42	│ │ ├── 66b428b9-55dc-4907-b116-55aaa887de30.png
43	│ │ ├── [...]
44	│ │ └── part-000001.json
45	│ ├── part-000002
46	│ ├── part-000003
47	│ ├── [...]
48	│ └── part-002000
49	└── metadata.parquet
50	```
51
52	```bash
53	# DiffusionDB Large
54	./
55	├── diffusiondb-large-part-1
56	│ ├── part-000001
57	│ │ ├── 0a8dc864-1616-4961-ac18-3fcdf76d3b08.webp
58	│ │ ├── 0a25cacb-5d91-4f27-b18a-bd423762f811.webp
59	│ │ ├── 0a52d584-4211-43a0-99ef-f5640ee2fc8c.webp
60	│ │ ├── [...]
61	│ │ └── part-000001.json
62	│ ├── part-000002
63	│ ├── part-000003
64	│ ├── [...]
65	│ └── part-010000
66	├── diffusiondb-large-part-2
67	│ ├── part-010001
68	│ │ ├── 0a68f671-3776-424c-91b6-c09a0dd6fc2d.webp
69	│ │ ├── 0a0756e9-1249-4fe2-a21a-12c43656c7a3.webp
70	│ │ ├── 0aa48f3d-f2d9-40a8-a800-c2c651ebba06.webp
71	│ │ ├── [...]
72	│ │ └── part-010001.json
73	│ ├── part-010002
74	│ ├── part-010003
75	│ ├── [...]
76	│ └── part-014000
77	└── metadata-large.parquet
78	```
79
80	These sub-folders have names `part-0xxxxx`, and each image has a unique name generated by [UUID Version 4](https://en.wikipedia.org/wiki/Universally_unique_identifier). The JSON file in a sub-folder has the same name as the sub-folder. Each image is a `PNG` file (DiffusionDB 2M)

People who liked this also viewed

6 picks

Type

diffusiondb

1.4kstars79forks

View on GitHub

How to install diffusiondb?

poloclub/diffusiondb

$git clone https://github.com/poloclub/diffusiondb

Installs into the current project.

Summary

Files · 1

View on GitHub

README.md

1	# DiffusionDB <a href="https://huggingface.co/datasets/poloclub/diffusiondb"><picture><source media="(prefers-color-scheme: dark)" srcset="https://i.imgur.com/yGxUUlX.png"><img src="favicon.ico" align="right" src="favicon.ico" height="40px"></picture>
2
3	[![hugging](https://img.shields.io/badge/🤗%20Hugging%20Face-Datasets-yellow)](https://huggingface.co/datasets/poloclub/diffusiondb)
4	[![license](https://img.shields.io/badge/License-CC0/MIT-blue)](#licensing)
5	[![arxiv badge](https://img.shields.io/badge/arXiv-2210.14896-red)](https://arxiv.org/abs/2210.14896)
6	[![datasheet](https://img.shields.io/badge/Data%20Sheet-Available-success)](https://poloclub.github.io/diffusiondb/datasheet.html)
7	<!-- [![DOI:10.1145/3491101.3519653](https://img.shields.io/badge/DOI-10.1145/3491101.3519653-blue)](https://doi.org/10.1145/3491101.3519653) -->
8
9	<img width="100%" src="https://user-images.githubusercontent.com/15007159/201762588-f24db2b8-dbb2-4a94-947b-7de393fc3d33.gif">
10
11	DiffusionDB is the first large-scale text-to-image prompt dataset. It contains 14 million images generated by Stable Diffusion using prompts and hyperparameters specified by real users. The unprecedented scale and diversity of this human-actuated dataset provide exciting research opportunities in understanding the interplay between prompts and generative models, detecting deepfakes, and designing human-AI interaction tools to help users more easily use these models.
12
13	## Get Started
14
15	DiffusionDB is available at [🤗 Hugging Face Datasets](https://huggingface.co/datasets/poloclub/diffusiondb).
16
17	## Two Subsets
18
19	DiffusionDB provides two subsets (DiffusionDB 2M and DiffusionDB Large) to support different needs.
20
21	\|Subset\|Num of Images\|Num of Unique Prompts\|Size\|Image Directory\|Metadata Table\|
22	\|:--\|--:\|--:\|--:\|--:\|--:\|
23	\|DiffusionDB 2M\|2M\|1.5M\|1.6TB\|`images/`\|`metadata.parquet`\|
24	\|DiffusionDB Large\|14M\|1.8M\|6.5TB\|`diffusiondb-large-part-1/` `diffusiondb-large-part-2/`\|`metadata-large.parquet`\|
25
26	##### Key Differences
27
28	1. Two subsets have a similar number of unique prompts, but DiffusionDB Large has much more images. DiffusionDB Large is a superset of DiffusionDB 2M.
29	2. Images in DiffusionDB 2M are stored in `png` format; images in DiffusionDB Large use a lossless `webp` format.
30
31	## Dataset Structure
32
33	We use a modularized file structure to distribute DiffusionDB. The 2 million images in DiffusionDB 2M are split into 2,000 folders, where each folder contains 1,000 images and a JSON file that links these 1,000 images to their prompts and hyperparameters. Similarly, the 14 million images in DiffusionDB Large are split into 14,000 folders.
34
35	```bash
36	# DiffusionDB 2M
37	./
38	├── images
39	│ ├── part-000001
40	│ │ ├── 3bfcd9cf-26ea-4303-bbe1-b095853f5360.png
41	│ │ ├── 5f47c66c-51d4-4f2c-a872-a68518f44adb.png
42	│ │ ├── 66b428b9-55dc-4907-b116-55aaa887de30.png
43	│ │ ├── [...]
44	│ │ └── part-000001.json
45	│ ├── part-000002
46	│ ├── part-000003
47	│ ├── [...]
48	│ └── part-002000
49	└── metadata.parquet
50	```
51
52	```bash
53	# DiffusionDB Large
54	./
55	├── diffusiondb-large-part-1
56	│ ├── part-000001
57	│ │ ├── 0a8dc864-1616-4961-ac18-3fcdf76d3b08.webp
58	│ │ ├── 0a25cacb-5d91-4f27-b18a-bd423762f811.webp
59	│ │ ├── 0a52d584-4211-43a0-99ef-f5640ee2fc8c.webp
60	│ │ ├── [...]
61	│ │ └── part-000001.json
62	│ ├── part-000002
63	│ ├── part-000003
64	│ ├── [...]
65	│ └── part-010000
66	├── diffusiondb-large-part-2
67	│ ├── part-010001
68	│ │ ├── 0a68f671-3776-424c-91b6-c09a0dd6fc2d.webp
69	│ │ ├── 0a0756e9-1249-4fe2-a21a-12c43656c7a3.webp
70	│ │ ├── 0aa48f3d-f2d9-40a8-a800-c2c651ebba06.webp
71	│ │ ├── [...]
72	│ │ └── part-010001.json
73	│ ├── part-010002
74	│ ├── part-010003
75	│ ├── [...]
76	│ └── part-014000
77	└── metadata-large.parquet
78	```
79
80	These sub-folders have names `part-0xxxxx`, and each image has a unique name generated by [UUID Version 4](https://en.wikipedia.org/wiki/Universally_unique_identifier). The JSON file in a sub-folder has the same name as the sub-folder. Each image is a `PNG` file (DiffusionDB 2M)

People who liked this also viewed

6 picks

Type