libtorch-yolov3

A Libtorch implementation of the YOLO v3 object detection algorithm, written with pure C++. It's fast, easy to be integrated to your production, and supports CPU and GPU computation. Enjoy ~

The code is based on the pytorch version, I rewritten it with C++.

Requirements

1. LibTorch latest build
2. Cuda
3. OpenCV (just used in the example)

To compile

1. cmake3
2. gcc 5.4 +

mkdir build && cd build
cmake3 -DCMAKE_PREFIX_PATH="your libtorch path" ..

# if there is multi versions of gcc, then tell cmake which one your want to use, e.g.:
cmake3 -DCMAKE_PREFIX_PATH="your libtorch path" -DCMAKE_C_COMPILER=/usr/local/bin/gcc -DCMAKE_CXX_COMPILER=/usr/local/bin/g++ ..

Ruuning the detector

The first thing you need to do is to get the weights file for v3:

cd models
wget https://pjreddie.com/media/files/yolov3.weights 

On Single image:

./yolo-app ../imgs/person.jpg