A modified version of Darknet with data and configuration to train a Yolov3 garbage object detection mode.
Darknet is an open source neural network framework written in C and CUDA. It is fast, easy to install, and supports CPU and GPU computation.
For more information see the Darknet project website.
For questions or issues please use the Google Group.
To download pretrained weights, images, labels and configurations go here: https://drive.google.com/drive/folders/1DjeNxdaF7AW3Nu54_3oRw_1SeYJtOvNL
One of the datasets contains images and annotations of cigarettes. The other data set has garbage bags, cardboard and containers.
First make Darknet. This website also has more indepth instructions.
To run training:
./darknet detector train cfg/garb.data cfg/yolov3_garb.cfg backup/yolov3_garb.backup > logs/train.log
To monitor loss during trainning, use jupyter notebook eval.pynb
For making predictions and test use this repository.