This is a collection of small(and possibly not) tools made by me for various specific tasks that i've encountered in my trainings, and thought i could speed up.

Most of them are part of this repo - https://github.com/Anzhc/Anzhc-s-Dataset-Processing-Tools/tree/main or standalone repos on my account.

Written by GPT4, as im not a programmer myself. Sometimes fixed and polished by @richyrich515 and @Jelosus1, so shoutouts to them!

Categorization GUI

Simple tool for categorizing datasets. Streamlines sorting and tagging of specific areas of concepts. Was used in @FallenIncursio's new UBW version, and released alongside it.

Helps with figuring out and streamlining process of tagging complex specific categories.

Create category boxes, load your image folder, look at first image, name first category, click on box! Your preview for that category will stay, and you will easily see what you have categorized.

Brightness Tagging Script

Small script that will add category of brightness for your image at the end in according text file. Grouping by brightness works pretty good, especially if you process large datasets.

Did not show significant change on XL, but likely will be useful for future SD3 and current sd1.5.

Fast Downscale

Helps with managing dataset size, especially large one. Downscales images to target resolution set in script, and converts it to 95 quality jpg. Multithreaded.