Library-Mutsumi/AnimeText_yolo

These are the YOLO12 models for our dataset AnimeText: A Large-scale Dataset for Robust Complex Anime Scene Text Detection.

You can try them in our online demo.

Our paper has been published at https://arxiv.org/abs/2510.07951

Model	Type	FLOPS	Params	F1 Score	Threshold	precision(B)	recall(B)	mAP50(B)	mAP50-95(B)	F1 Plot	Confusion	Labels
yolo12x_animetext	yolo	200G	59.1M	0.9	0.425	0.89785	0.89292	0.95167	0.89656	plot	confusion	text_block
yolo12l_animetext	yolo	89.4G	26.4M	0.89	0.426	0.89415	0.8906	0.94905	0.88587	plot	confusion	text_block
yolo12m_animetext	yolo	67.7G	20.1M	0.89	0.299	0.88725	0.88466	0.94556	0.87804	plot	confusion	text_block
yolo12n_animetext	yolo	6.48G	2.57M	0.86	0.251	0.8649	0.85504	0.9231	0.83449	plot	confusion	text_block
yolo12s_animetext	yolo	21.5G	9.25M	0.88	0.272	0.87922	0.87581	0.93891	0.86523	plot	confusion	text_block