I have successfully converted darknet pretrained model into tflite model using tensorflow 2.3.0 or rc version and opencv 4.1.1.26. Then, I ran detect.py but the scales of bounding boxes do not match. Any suggestions? I have fixed configuration paramters for anchor sizes, input sizes and all I believe. If I have missed anything, please let me know.
Hello,
I have successfully converted darknet pretrained model into tflite model using tensorflow 2.3.0 or rc version and opencv 4.1.1.26. Then, I ran detect.py but the scales of bounding boxes do not match. Any suggestions? I have fixed configuration paramters for anchor sizes, input sizes and all I believe. If I have missed anything, please let me know.