ptx0/free-to-use-pixelart

---
license: mit
---

# Free-to-use Pixel Art

## Dataset Details

This dataset was collected on 25th May, 2024.

It's a small subset of the free-to-use images on PixilArt.

At the time of publication, this dataset was covered by permissive terms that allow commercial use.

### Dataset Description

This dataset is unique in that it contains the pixel group size for each collected sample, which might assist in experiments on microconditioning inputs on an adapter to control this value of the unit more precisely.

- **Curated by:** @pseudoterminalx
- **License:** Free-to-use, commercial license

### Dataset Sources

- **Repository:** https://pixilart.com

## Dataset Structure

A single parquet file contains several thousand rows, each row is a single sample. The dataset has been deduplicated and filtered.

The elements were extracted and transformed from the API via this code:

```py
'subset': f"{subset}",
'sequence': f"{sequence}",
'subset-sequence-element': f"{subset}.{sequence}.{element_idx}",
'title': element['title'],
'description': element['description'],
'filename': f"{element['image_id']}-{element_idx} {element['title']}.png",
'views': element['views'],
'image_hash': element['image_id'],
'image_url': element['image_url'],
'full_image_url': element['full_image_url'],
'likes_count': element['likes_count'],
'pixel_size': element.get('pixel_size', 0),
'has_watermark': element.get('has_watermark', False),
'comments_count': element['comments_count'],
'width': element['width'],
'height': element['height'],
'date_created': element['date_created'],
'content_warning': element['content_warning'] if 'content_warning' in element else None,
'warning': str(element['warning']),
'liked': element['liked']
```

Leading to the following fields in the dataset:

- subset (str) - the subset of the API this was retrieved from, eg. highlighted or staff-picks
- sequence (int) - a numeric index that is sent to the API to retrieve the next results
- subset-sequence-element (str) - a combined string used to index the dataset for unique entries
- title (str) - the original post title as returned via the API
- description (str) - the original post description
- filename (str) - a suggested filename to use for downloading the URL should you directly use this in a training tool
- views (int) - how many views this sample has
- image_hash (str) - not useful for validating image contents
- image_url (str) - a thumbnail of the image
- full_image_url (str) - the original, full-size image
- likes_count (int) - how many likes the post had
- pixel_size (int) - the pixel group size for the image, with lower values being more like simple MS Paint drawings
- has_watermark (bool) - the value returned by the API (defaulting to False if unavailable) telling us whether the image is watermarked
- comments_count (int) - how many comments the post had
- width, height (int) - the image dimensions
- content_warning (bool) - whether the content is deemed sensitive or potentially offensive by the vendor
- warning (str) - the contents of the warning text, explaining the potential problem
- liked (bool) - whether the image has been liked

原始信息汇总

数据集概述

基本信息

• 名称: Free-to-use Pixel Art
• 收集日期: 2024年5月25日
• 来源: PixilArt
• 许可证: 免费使用，商业许可证
• 策划者: @pseudoterminalx

数据集描述

• 该数据集包含每个收集样本的像素组大小，有助于在适配器上进行微调输入，以更精确地控制单元值。

数据集结构

• 数据集存储为单个parquet文件，包含数千行，每行代表一个单独样本。
• 数据集已去重和过滤。
• 数据集包含以下字段：
  • subset (str)
  • sequence (int)
  • subset-sequence-element (str)
  • title (str)
  • description (str)
  • filename (str)
  • views (int)
  • image_hash (str)
  • image_url (str)
  • full_image_url (str)
  • likes_count (int)
  • pixel_size (int)
  • has_watermark (bool)
  • comments_count (int)
  • width, height (int)
  • content_warning (bool)
  • warning (str)
  • liked (bool)