This dataset is designed to train models for monochrome image classification, and the data is mainly sourced from danbooru and pixiv.
All images in this dataset are guaranteed to be in JPEG format with a file extension of .jpg.
Currently, it has three different versions:
Version
Description
Monochrome Images
Normal Images
Total Images
v1.0
Oldest version
304
174
478
v1.1
Balanced samples
528
514
1042
v2.0
More complex samples
2208
2250
4458
