Tuxemon Dataset
This dataset contains images of mosnters from The Tuxemon Project - an open source effort for a monster catching game.
These image-caption pairs can be used for text-to-image tuning & benchmarking.
All images in this dataset were downloaded from
https://wiki.tuxemon.org/Category:Monster
Some images were upscaled using SDx4 upscaler & HiDiffusion
Captions generated with
BLIP-large (some were manually modified)
GPT-4 Turbo
[!TIP]
One can use the mix of captions provided in the… See the full description on the dataset page:
https://huggingface.co/datasets/diffusers/tuxemon.
