Closed makra89 closed 2 years ago
See comment in See https://github.com/zzh8829/yolov3-tf2/pull/311
I think the score in yolo_nms is calculated wrong when only having one class. The model cannot learn anything about the class since the associated loss is always 0 --> class probability is always <= 0.5 --> score is always <= 0.5
Ah and by the way: Awesome work :) Your model works very well when trained trying to build an aimbot: https://github.com/makra89/Quake_AI
I don't know if it's commonplace to write on these, but with this change my one class object detector works perfectly thanks!
See comment in See https://github.com/zzh8829/yolov3-tf2/pull/311
I think the score in yolo_nms is calculated wrong when only having one class. The model cannot learn anything about the class since the associated loss is always 0 --> class probability is always <= 0.5 --> score is always <= 0.5
Ah and by the way: Awesome work :) Your model works very well when trained trying to build an aimbot: https://github.com/makra89/Quake_AI