xuanandsix / visdrone-yolov8

Many yolov8 model are trained on the VisDrone dataset.

VisDrone-yolov8

Many yolov8 models are trained on the VisDrone dataset. Office code https://github.com/ultralytics/ultralytics.

Train

You can train according to my parameter settings or download my model. All model parameters were trained on a single A40 GPU. Using yolov8s as main comparison baseline.

Performance

model	input size	other hyper-parameters	val mAP50	val mAP50-95	test mAP50	test mAP50-95	model & log
yolov8s	640	default	0.404	0.24	0.32	0.184	Google
yolov8l	640	default	0.453	0.278	0.361	0.213	Google
yolov8s	960	default	0.496	0.308	0.396	0.234	Google
yolov8s	960 & mutil-scale	default	0.492	0.304	0.399	0.236	Google
yolov8s-p2	960	default	0.519	0.321	0.417	0.244	Google
yolov8s-p2	960	minxup=0.4	0.522	0.325	0.431	0.255	Google
yolov8s-p2	960	minxup=0.4, val iou=0.6	0.531	0.327	0.437	0.256	Google

The parameters settings for best performance can be found in the default.yml file.

Other failed attempts

• copy_paste
• reduce image scale (+/- gain)

visdrone-yolov8's Issues

P2

what's meaning of “p2”？