Lightnet documentation

Lightnet is a library for pytorch, that makes it easier to create CNN’s. It was mainly created to implement darknet networks in python, but has since grown into a more general purpose single-shot object detection network.

Credits

Credits where credits are due. I get a lot -if not most- of my ideas from other papers and repositories.
Without the work of the following people, this library would have never become a reality.

  • pjreddie et al. for their work on the original YOLO networks

  • The PyTorch team for creating a clear and powerfull library for deep learning.

  • marvis for his implementation of YOLO in pytorch. I took a lot of parts from this repository, and tweaked them to my needs.

  • longcw for his implementation of YOLO in pytorch. I used his code to get a correct region loss.

  • All authors who’s papers I used to develop my research.

  • The contributors that helped push Lightnet to the library it is today!

Cite

If you use Lightnet in your research, please cite it.

@misc{lightnet18,
  author = {Tanguy Ophoff},
  title = {Lightnet: Building Blocks to Recreate Darknet Networks in Pytorch},
  howpublished = {\url{https://gitlab.com/EAVISE/lightnet}},
  year = {2018}
}

Table of Contents

C++

Indices and tables