ivonatau / python-wrapper-for-yolo

Python bindings for YOLO v2 (https://pjreddie.com/darknet/yolo/) with ctypes.

Python-wrapper-for-YOLO

Python bindings for YOLO v2 (https://pjreddie.com/darknet/yolo/) detector function with ctypes.

Install perequisites.

• Command line tools (gcc)
• PIL

pip install pillow

Download YOLOv2 source code

git clone https://github.com/pjreddie/darknet
cd darknet

Replace makefile with the one from this repository and run make

make
wget http://pjreddie.com/media/files/yolo.weights

Create shared library object and copy it to your darknet directory

cd src
gcc -c -fPIC *.c
gcc -shared -fPIC  -o libdarknet.so  *.o -lc
cp ./libdarknet.so ../
cd ..

Copy python_wrapper.py from this repository to your darknet directory

Put all your test files to darknet/data directory and run the script.

python python_wrapper.py

python-wrapper-for-yolo's Issues

Some questions ...

Hello,

I appreciate your reply for these important questions:

1. Does your code support GPU (CUDA, CuDNN) for the maximum acceleration?
2. in the code you mentioned:

mylib = ctypes.cdll.LoadLibrary('libdarknet.so')

Where should I find the libdarknet.so file?

3. is it possible to test the code with a single image?
4. How can we feed the code with video from the camera?
5. Does your code draw bounding boxes over detected objects and handles the coordinates?