search

cwq159 / PyTorch-Spiking-YOLOv3

A PyTorch implementation of Spiking-YOLOv3. Two branches are provided, based on two common PyTorch implementation of YOLOv3(ultralytics/yolov3 & eriklindernoren/PyTorch-YOLOv3), with support for Spiking-YOLOv3-Tiny at present.
GNU General Public License v3.0
207 stars 58 forks source link

Is it possible to fine-tune using Darknet weights #29

Open JNaranjo-Alcazar opened 3 years ago

JNaranjo-Alcazar commented 3 years ago

Hi community, first of all, thanks for this amazing contribution.

I was wondering if it is possible to fine tune a pre-trained yolo network, for that I assume that --weights parameter should point to a downloaded file, in my case darknet53.conv.74 and then use --freze parameter and set it to True.

Would the conversion to spiking be run properly?

Thanks in advance!

cwq159 commented 3 years ago

Sorry, it is impossible to fine-tune directly using Darknet weights. Because some operators(layers) can't be supported for now. You have to convert them first and train a converted pre-train model.