id
stringlengths
10
13
original_prompt
stringlengths
7
3.51k
positive_prompt
stringlengths
20
3.52k
negative_prompt
stringlengths
0
79
model
stringclasses
1 value
filepath
stringlengths
26
29
num_inference_steps
int64
40
200
width
int64
512
512
height
int64
512
512
url
stringlengths
25
1.81k
image
imagewidth (px)
512
512
heatmap_labels
sequencelengths
0
50
heatmaps
sequencelengths
0
50
1042658020996
Series 1 Meaningful Use for Behavioral Health Providers
An image of Series 1 Meaningful Use for Behavioral Health Providers
stabilityai/stable-diffusion-2-1-base
part-000090/1042658020996.jpg
180
512
512
http://thumbs.slideserve.com/1_2509937.jpg
[ "series", "meaningful", "use", "behavioral", "health", "providers" ]
[[[0.291015625,0.28564453125,0.278564453125,0.27734375,0.297607421875,0.29248046875,0.29638671875,0.(...TRUNCATED)
1042658022076
Frog LED Flashlight Keychain with Froggy Sound Effects
"A high quality picture of Frog LED Flashlight Keychain with Froggy Sound Effects, Medium shot, Sun(...TRUNCATED)
low-resolution, poorly drawn face, render, bad anatomy, retro
stabilityai/stable-diffusion-2-1-base
part-000090/1042658022076.jpg
60
512
512
http://miniimg8.rightinthebox.com/images/190x176/201207/gdlayb1343616636269.jpg
[ "frog", "led", "flashlight", "sound", "effects" ]
[[[0.140625,0.1575927734375,0.1898193359375,0.190185546875,0.196533203125,0.196533203125,0.202392578(...TRUNCATED)
1042658022246
Sharukh Khan... the cheesiest actor EVER
An image of Sharukh Khan... the cheesiest actor EVER
stabilityai/stable-diffusion-2-1-base
part-000090/1042658022246.jpg
40
512
512
https://i.pinimg.com/236…6a4431bb596d.jpg
[ "khan", "cheesiest", "actor", "ever" ]
[[[0.056610107421875,0.054901123046875,0.0582275390625,0.05865478515625,0.0589599609375,0.0671386718(...TRUNCATED)
1042658026540
Image for 'The Marshall Tucker Band'
"A photorealistic photo of Image for 'The Marshall Tucker Band', POV, Soft light, At night, Teleph(...TRUNCATED)
stabilityai/stable-diffusion-2-1-base
part-000090/1042658026540.jpg
50
512
512
http://img2-ak.lst.fm/i/u/300x300/8ff3b4e7b69641e79a86749e5a408cda.jpg
[ "image", "marshall", "tucker", "band" ]
[[[0.1114501953125,0.1024169921875,0.10394287109375,0.106689453125,0.11370849609375,0.1153564453125,(...TRUNCATED)
1042658026677
Strato Bathroom Furniture Set 15 Meubles souslavabo Inbani
"A dramatic picture of Strato Bathroom Furniture Set 15 Meubles souslavabo Inbani, Long shot, In(...TRUNCATED)
out of frame, fuzzy, mutation, digital art, bad anatomy
stabilityai/stable-diffusion-2-1-base
part-000090/1042658026677.jpg
50
512
512
https://image.architonic…-set-05-2-sq.jpg
[ "bathroom", "furniture", "set" ]
[[[0.3935546875,0.410888671875,0.436279296875,0.472900390625,0.51904296875,0.54248046875,0.591796875(...TRUNCATED)
1042658027470
Frosty Jacks White50 Extra FreeBig Value Pack
An image of Frosty Jacks White50 Extra FreeBig Value Pack
low-resolution, grainy, bad art, disfigured, cartoon
stabilityai/stable-diffusion-2-1-base
part-000090/1042658027470.jpg
60
512
512
https://www.bestwaywhole…012136007308.jpg
[ "frosty", "jacks", "extra", "value", "pack" ]
[[[0.4560546875,0.4873046875,0.45947265625,0.43994140625,0.411865234375,0.418212890625,0.40234375,0.(...TRUNCATED)
1042658031860
Steering Gear tipe Recirculating Ball dan Rack and Pinion
"A hyper realistic picture of Steering Gear tipe Recirculating Ball dan Rack and Pinion, Medium sh(...TRUNCATED)
colorless, misshaped, graphic, computer graphic, digital art
stabilityai/stable-diffusion-2-1-base
part-000090/1042658031860.png
50
512
512
http://1.bp.blogspot.com/-LsqZDL6DXsM/VItpFZVMlzI/AAAAAAAABG8/dGXlATln3Gk/s1600/steering-gear.png
[ "steering", "gear", "ball", "rack", "pinion" ]
[[[0.050445556640625,0.05926513671875,0.068603515625,0.076904296875,0.08544921875,0.09356689453125,0(...TRUNCATED)
1042658033302
Cat Coming Home A Joe Grey Mystery, by Shirley Rousseau Murphy
An image of Cat Coming Home A Joe Grey Mystery, by Shirley Rousseau Murphy
stabilityai/stable-diffusion-2-1-base
part-000090/1042658033302.jpg
50
512
512
https://s3-us-west-2.ama…r-square-175.jpg
[ "cat", "coming", "home", "a", "grey", "mystery", "rousseau", "murphy" ]
[[[0.14404296875,0.181884765625,0.1806640625,0.19970703125,0.1898193359375,0.1824951171875,0.2111816(...TRUNCATED)
1042659001822
LINZI Blue White Polka Dot Bow T Bar Very High Stiletto Heel Party Shoes Size 3
An image of LINZI Blue White Polka Dot Bow T Bar Very High Stiletto Heel Party Shoes Size 3
stabilityai/stable-diffusion-2-1-base
part-000090/1042659001822.jpg
50
512
512
http://thumbs.ebaystatic.com/images/g/yp8AAOSwLF1X5Vu3/s-l225.jpg
["blue","white","polka","dot","bow","t","bar","very","high","stiletto","heel","party","shoes","size"(...TRUNCATED)
[[[0.14501953125,0.151611328125,0.1484375,0.1463623046875,0.1505126953125,0.156494140625,0.155029296(...TRUNCATED)
1042659006875
Asura's Wrath PS3
An image of Asura's Wrath PS3
stabilityai/stable-diffusion-2-1-base
part-000090/1042659006875.jpg
50
512
512
https://www.gamehope.com…h.jpg?1432126264
[ "asura", "wrath" ]
[[[0.318359375,0.33251953125,0.35986328125,0.336669921875,0.337158203125,0.326416015625,0.3330078125(...TRUNCATED)

ELSA - Multimedia use case

daam.gif

ELSA Multimedia is a large collection of Deep Fake images, generated using diffusion models

Dataset Summary

This dataset was developed as part of the EU project ELSA. Specifically for the Multimedia use-case. Official webpage: https://benchmarks.elsa-ai.eu/ This dataset aims to develop effective solutions for detecting and mitigating the spread of deep fake images in multimedia content. Deep fake images, which are highly realistic and deceptive manipulations, pose significant risks to privacy, security, and trust in digital media. This dataset can be used to train robust and accurate models that can identify and flag instances of deep fake images.

ELSA versions

Name Description Link
ELSA1M_track1 Dataset of 1M images generated using diffusion model https://huggingface.co/datasets/elsaEU/ELSA1M_track1
ELSA500k_track2 Dataset of 500k images generated using diffusion model with diffusion attentive attribution maps [1] https://huggingface.co/datasets/elsaEU/ELSA500k_track2
from daam import WordHeatMap
from datasets import load_dataset
import torch

elsa_data = load_dataset("elsaEU/ELSA500k_track2", split="train", streaming=True)
for sample in elsa_data:
  image = sample.pop("image")
  heatmaps = sample.pop("heatmaps")
  heatmap_labels = sample.pop("heatmap_labels")
  metadata = sample
  for j, (h, l) in enumerate(zip(heatmaps, heatmap_labels)):
    heatmap = WordHeatMap(torch.Tensor(h), word=l)
    heatmap.plot_overlay(image)
    plt.show()

Using streaming=True lets you work with the dataset without downloading it.

Dataset Structure

Each parquet file contains nearly 1k images and a JSON file with metadata.

The Metadata for generated images are:

  • ID: Laion image ID
  • original_prompt: Laion Prompt
  • positive_prompt: positive prompt used for image generation
  • negative_prompt: negative prompt used for image generation
  • model: model used for the image generation
  • nsfw: nsfw tag from Laion
  • url_real_image: Url of the real image associated to the same prompt
  • filepath: filepath of the fake image
  • aspect_ratio: aspect ratio of the generated image
  • heatmaps: diffusion attentive attribution maps
  • heatmap_labels: words releated to the heatmaps

Dataset Curators

References

[1] What the DAAM: Interpreting Stable Diffusion Using Cross Attention, 2023. Tang Raphael et al.

Downloads last month
357