Is is possible to load a binary file which was saved in darknet framework?

Hello guys,
I have got a question. I would like to reimplement the yolo v2 in pytorch using pretrained weight which was released in pjreddie homepage with this link. The pretrained weights are in binary format! I would like to know if I have the architecture of yolov2, Is it possible to use these weights for loading them into pytorch model?

it’s certainly possible, but no one has written any such code, so you might have to figure it out yourself on how to translate their binary format into a torch Tensor or numpy array.

1 Like

Could you please give me some hints about how can I start to do that?

I believe this implementation includes something that reads Darknet weights:


Thanks :slight_smile: is better.

1 Like