AlexeyAB / darknet

YOLOv4 / Scaled-YOLOv4 / YOLO - Neural Networks for Object Detection (Windows and Linux version of Darknet )
http://pjreddie.com/darknet/
Other
21.75k stars 7.96k forks source link

No detection is found after training #2370

Open segsev opened 5 years ago

segsev commented 5 years ago

I am using yolov3 tiny for detecting the boundary box of an object. I trained over 14000 labelled images. And while predicting the boundary box on images, I am getting 0,0,0,0 as the boundary box co ordinates for few images. After further investigating, I found that the key "detections" is returning an empty list for those images. Is there any way to fix it out?

detections = { "detections": detections, "image": image, "caption": "\n
".join(imcaption) }