I tried some strategies to optimize the yolov3 network structure, removing some layers and channels.
When I use mystic123/convert_weights_pb.py， error happened：ValueError: cannot reshape array of size 27240 into shape (256,128,3,3)
so if i changed .cfg，is there a method i can convert my .weights to .pb and IR.And can openvino inferences accept changed yolo.cfg
Besides,when will openvino support yolov4 ^-^
After you optimized the yoloV3 network structure, what command did you use to convert to IR format?
Here are the instructions for converting your .weights file and then converting to IR.