search

allanzelener / YAD2K

YAD2K: Yet Another Darknet 2 Keras
Other
2.72k stars 881 forks source link

yolo_loss #160

Open jjjj4444 opened 5 years ago

jjjj4444 commented 5 years ago

in yolo_loss function when calculateno_object_weights = (no_object_scale * (1 - object_detections) *(1 - detectors_mask)) should no_object_weights = (no_object_scale * (object_detections) *(1 - detectors_mask)) since no_object_weights is only activated when the iou between predicted boxes and ground truth is larger than 0.6 and detectors_mask=0.

speeding-motor commented 4 years ago

I have a little confuse with the loss function:

why not only use detector_mask to judge the box have box or not, why still need to get the object_detections,

any help is appreciate ,thanks in advance,