Closed kalikhademi closed 3 years ago
Hi,
I am training yolo version 3 on tensorflow to detect water towers on pictures. I have followed all the instruction s in this link: https://github.com/zzh8829/yolov3-tf2/blob/master/docs/training_voc.md
I trained the model using converted darknet weights. The training goes well without any errors. When I use detect.py on the tensorflow checkpoints, all the classes, score and boxes
are tensors with zero values and no detection is shown.
What could be the problem about this issue?
I made the change that is suggested in issue #70 and it solved the training problem. I also trained the classifier using darknet weights in eager-fit.
Hi,
I am training yolo version 3 on tensorflow to detect water towers on pictures. I have followed all the instruction s in this link: https://github.com/zzh8829/yolov3-tf2/blob/master/docs/training_voc.md
I trained the model using converted darknet weights. The training goes well without any errors. When I use detect.py on the tensorflow checkpoints, all the classes, score and boxes
are tensors with zero values and no detection is shown.
What could be the problem about this issue?