YOLO Object Detection¶
1. Set up your environment.¶
export CUDA_HOME=/usr/local/cuda
export PATH=$CUDA_HOME/bin:$PATH
export LD_LIBRARY_PATH=$CUDA_HOME/lib64:$LD_LIBRARY_PATH
Note
You can also add path to .bashrc.
3. Retrive darknet repository from github.¶
sudo apt update
sudo apt install git
git clone https://github.com/AlexeyAB/darknet.git
4. Build darknet with CUDNN and OpenCV support.¶
Modify darknet’s Makefile with the following:
GPU=1
CUDNN=1
OPENCV=1
sudo apt update
sudo apt install make build-essential
cd darknet
# Edit Makefile GPU=1, CUDNN=1, OPENCV=1
make
5. Download the pre-trained weights.¶
wget https://pjreddie.com/media/files/yolov3-tiny.weights
Note
ADLINK doesn’t own the pre-trained data. This pre-trained data is a contribution from original author in community.
6. Run object detector from example image.¶
./darknet detect cfg/yolov3-tiny.cfg yolov3-tiny.weights data/dog.jpg
