Datasets:
File size: 3,305 Bytes
11d4be5 2bc9bbb 351c8cf f67e9d9 2bc9bbb f67e9d9 2bc9bbb 22647c8 2bc9bbb 20a3645 2bc9bbb f67e9d9 2bc9bbb f67e9d9 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 |
---
license: cc-by-nc-sa-4.0
task_categories:
- text-to-image
language:
- en
tags:
- art
- furry
- realism
- artstation
- url
- image
- text
- csv
pretty_name: Furry Aesthetic Realism Annotated Dataset
size_categories:
- 10K<n<100K
---
# FARAD - Furry Aesthetic Realism Annotated Dataset
The FARAD text to image dataset consists of ~18k publicly-available image URL-text pairs from the professional art portfolio website [Artstation](artstation.com).
The images were drawn from a set of ~600k furry-adjacent tagged images and then assessed for quality and content with [FARAMIR](https://github.com/redhottensors/FARAMIR) and collaboratively relabeled with simplified [e621](https://e621.net/) tags by [JTP-PILOT²](https://huggingface.co/RedRocket/JointTaggerProject).
## Version 1
Download Link: https://huggingface.co/datasets/RedRocket/FARAD/resolve/main/v1/data.csv?download=true (5MB)<br>
Statistics: 18409 images from 4472 distinct projects. 738 artists and 1937 distinct tags.<br>
CSV Fields: <code>url,project,asset,user,width,height,date,likes,title,labels</code>
## Dataset Sample
<table>
<tr>
<td><b>"Toothless - How To Train Your Dragon"</b> – donchuan3d</td>
<td><b>"Wolf Portrait (different angle)"</b> – massimorighi</td>
</tr>
<tr>
<td><img style="height: 386px; margin: 0;" src="https://cdna.artstation.com/p/assets/images/images/058/763/682/large/juan-hernandez-toothless-studio-1001.jpg"></td>
<td><img style="height: 386px; margin: 0;" src="https://cdnb.artstation.com/p/assets/images/images/016/724/971/large/massimo-righi-wolffrontmain.jpg"></td>
</tr>
<tr>
<td><code>3d, how to train your dragon, male feral, night fury, rigging, simple background, solo, stylized, substance painter, toothless, zbrush</code></td>
<td><code>3d, ambiguous feral, arnold, realistic, simple background, solo, wolf</td>
</tr>
<tr>
<td><b>"Jaarlandtir Harbour - Tech Art - Harbour Sea Water Shader..."</b> – khael</td>
<td><b>"Raptors hunting"</b> – alessandromastronardi</td>
</tr>
<tr>
<td><img style="height: 386px; margin: 0;" src="https://cdnb.artstation.com/p/assets/images/images/054/990/295/large/kevin-babin-artblast-harbour-water.jpg"></td>
<td><img style="height: 386px; margin: 0;" src="https://cdna.artstation.com/p/assets/images/images/006/836/146/large/alessandro-mastronardi-deinonychus-popup105.jpg"></td>
</tr>
<tr>
<td><code>3d, castle, cliff, detailed background, fog, landscape, realistic, rock, scenery, sea, zero pictured</td>
<td><code>3d, ambiguous feral, deinonychus, detailed background, duo, realistic, zbrush</code></td>
</tr>
</table>
## Notes
Per the Artstation terms of service, the dataset does not contain images tagged with <code>#NoAI</code> at the time of retrieval.<br>
Images of untextured models are present in the dataset, with the label <code>unfinished</code>, as assessed by FARAMIR.<br>
The dataset may contain small amounts of unlabelled adult or mature content, but efforts have been made to avoid this.
## License
This dataset is available under the terms of the [CC-BY-NC-SA 4.0](https://creativecommons.org/licenses/by-nc-sa/4.0/deed.en) license.<br>
This dataset only contains hyperlinks to publicly-available images and associated metadata.
You are responsible for obeying all applicable laws and terms of service when using this dataset. |