datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
pixel-art-normal-maps
Pixel-Art Normal Maps
Synthetic (pixel-art sprite → camera-space normal map) image pairs for training and
evaluating normal-map generation models for 2D pixel-art.
This dataset accompanies a master's thesis on normal-map generation for pixelated images.
Code: https://github.com/Dmiktor/pixel-art-normal-maps
Load
from datasets import load_dataset
ds = load_dataset("<HF_USERNAME>/pixel-art-normal-maps", split="train")
ex = ds[0]
ex["color"] # PIL.Image - RGBA… See the full description on the dataset page: https://huggingface.co/datasets/DmytroKhitro/pixel-art-normal-maps.mentis-fusion360-normal-map-publicsmall-normal-map-dataset