search

gudovskiy / yoloNCS

YOLO object detector for Movidius Neural Compute Stick (NCS)
MIT License
181 stars 78 forks source link

How are the pretrained weights generated? How do I generate them? #6

Open codingquark opened 6 years ago

codingquark commented 6 years ago

Are the pretrained weights generated by using create_yolo_caffemodel.py?

If I want to train the model on my own what would the process be? Should I first train yolo by darknet and then use create_yolo_caffemodel.py and create_yolo_prototxt.py?

gudovskiy commented 6 years ago

@codingquark I used pretrained weights. You can either train using darknet and then convert with the scripts or train in caffe directly.

codingquark commented 6 years ago

Training in caffe sounds the easiest approach, since I can avoid having to convert to caffe. I have been playing with https://github.com/xingwangsfu/caffe-yolo and it is not working for many .cfgs.