AlexeyAB / darknet

YOLOv4 / Scaled-YOLOv4 / YOLO - Neural Networks for Object Detection (Windows and Linux version of Darknet )
http://pjreddie.com/darknet/
Other
21.65k stars 7.96k forks source link

How much can we change width and height in .cfg files? #7968

Open BS-98 opened 3 years ago

BS-98 commented 3 years ago

Hi,

I want to use yolov4-csp model for inference on my dataset (people dataset). So I use yolov4-csp.cfg and yolov4-csp.weights files, but as far as I know these weights was obtained by training on 512x512 or 640x640 resolution. I see that I can run inference with changed width=1056and height=1056, but I dont know if I get more accurate detections with these options. Maybe I should run inference on 640x640 with that model (yolov4-csp.cfg and yolov4-csp.weights) ?

akashAD98 commented 3 years ago

I think you can do the experiment & check which one is best.