nyanko7/yandere2023

---
license: mit
task_categories:
- image-classification
- image-to-image
- text-to-image
language:
- en
- ja
pretty_name: yandere2023
size_categories:
- 1M<n<10M
---

# Yandere2023: A Large-Scale Crowdsourced and Tagged Anime Illustration Dataset

<!-- yande.re -->

Yandere2023 is a comprehensive anime image dataset with over 1.2 million high-quality images sourced from various materials, including key frames, manga scans, artbooks, and more. While the average number of tags per image is relatively low, the dataset boasts a diverse collection of images with exceptional quality.

- **Shared by:** Nyanko Devs
- **Language(s):** English, Japanese
- **License:** MIT

## Uses

## Format

The goal of the dataset is to be as easy as possible to use immediately, avoiding obscure file formats, while allowing simultaneous research & seeding of the torrent, with easy updates.

Images are provided in the full original form (be that JPG, PNG, GIF or otherwise) for reference/archival purposes, and bucketed into 1000 subdirectories 0000–0999 (0-padded), which is the Yandere ID modulo 1000 (ie. all images in 0999/ have an ID ending in ‘999’); IDs can be turned into paths by dividing & padding (eg. in Bash, BUCKET=$(printf "%04d" $(( ID % 1000 )) )) and then the file is at original/$BUCKET/$ID.$EXT.

The reason for the bucketing is that a single directory would cause pathological filesystem performance, and modulo ID is a simple hash which spreads images evenly without requiring additional future directories to be made or a filesystem IO to check where the file is. The ID is not zero-padded and files end in the relevant extension, hence the file layout looks like this:

```bash
$ tree / | less

/
├── yandere2023 -> /mnt/diffusionstorage/workspace/yandere/
│ ├── metadata
│ ├── readme.md
│ ├── original
│ │ ├── 0000 -> data-0000.tar
│ │ ├── 0001 -> data-0001.tar
│ │ │ ├── 10001.jpg
│ │ │ ├── 210001.png
│ │ │ ├── 3120001.webp
│ │ │ ├── 6513001.jpg

```

原始信息汇总

Yandere2023: A Large-Scale Crowdsourced and Tagged Anime Illustration Dataset

Yandere2023 是一个全面的动漫图像数据集，包含超过 120 万张高质量图像，来源包括关键帧、漫画扫描、艺术书籍等。尽管每张图像的平均标签数量相对较低，但该数据集以其多样性和图像质量而著称。

• 共享者： Nyanko Devs
• 语言： 英语、日语
• 许可证： MIT

用途

该数据集的目标是尽可能易于立即使用，避免使用晦涩的文件格式，同时允许同时进行研究和种子下载，并易于更新。

格式

图像以原始形式（JPG、PNG、GIF 等）提供，用于参考/存档目的，并分桶到 1000 个子目录 0000–0999（0 填充），这是 Yandere ID 模 1000（即 0999/ 中的所有图像的 ID 以 999 结尾）；ID 可以通过除法和填充转换为路径（例如在 Bash 中，BUCKET=$(printf "%04d" $(( ID % 1000 ))），然后文件位于 original/$BUCKET/$ID.$EXT。

分桶的原因是单个目录会导致文件系统性能问题，而模 ID 是一个简单的哈希，可以均匀分布图像，而无需创建额外的未来目录或检查文件位置的文件系统 IO。ID 不是零填充的，文件以相关扩展名结尾，因此文件布局如下：

bash $ tree / | less

/ ├── yandere2023 -> /mnt/diffusionstorage/workspace/yandere/ │ ├── metadata │ ├── readme.md │ ├── original │ │ ├── 0000 -> data-0000.tar │ │ ├── 0001 -> data-0001.tar │ │ │ ├── 10001.jpg │ │ │ ├── 210001.png │ │ │ ├── 3120001.webp │ │ │ ├── 6513001.jpg