AlexeyAB / darknet

YOLOv4 / Scaled-YOLOv4 / YOLO - Neural Networks for Object Detection (Windows and Linux version of Darknet )
http://pjreddie.com/darknet/
Other
21.77k stars 7.96k forks source link

Multilabel Classification in YOLOv4 #8284

Open audrey-siqueira opened 2 years ago

audrey-siqueira commented 2 years ago

Hi, I've been working with Yolov4 lately and I have the following question:

I trained yolov4 with the following classes:

[Ferrarri, Mclaren, Red, White, Black]

Is there any way yolo can return me to the most likely car class followed by the most likely color class?

Of course I could train with the following classes: [Ferrarri_Red, Mclaren_Red, Ferrarri_White, Mclaren_White, Ferrarri_Black, Mclaren_Black]

The second option seems to be more accurate but it would cause me more work to make the labels manually, so I would like to know if there is any way to use the first option ?

Thanks

audrey-siqueira commented 2 years ago

Basically this is the same doubt : https://github.com/AlexeyAB/darknet/issues/7906

Could you help me @AlexeyAB ?

lsd1994 commented 2 years ago

[Ferrarri_Red, Mclaren_Red, Ferrarri_White, Mclaren_White, Ferrarri_Black, Mclaren_Black] Only this option is right.