Skip to content

Latest commit

 

History

History
23 lines (19 loc) · 1.79 KB

README.md

File metadata and controls

23 lines (19 loc) · 1.79 KB
Raw

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)