How to initial weights when training Single Class?

ultralytics / yolov3

YOLOv3 in PyTorch > ONNX > CoreML > TFLite

https://docs.ultralytics.com

GNU Affero General Public License v3.0

10.25k stars 3.45k forks source link

How to initial weights when training Single Class? #518

Closed QianSunCN closed 5 years ago

QianSunCN commented 5 years ago

Hello, your code is very beautyful. I want to train Single Class, and following the "Example: Train Single Class" , it works. There is one question:
How to initial the weights ?
Is there some pretrained model for yolov3-spp-1cls.cfg?

glenn-jocher commented 5 years ago

@QianSunCN you can initialize with random weights, darknet53.conv.74 backbone, or yolov3-spp:

--weights ''
--weights weights/darknet53.conv.74
--weights weights/yolov3-spp.pt