yolo v2 install
First, using Windows 7 to achieve
1. vs2015+opencv3.2+GeForce GT 1030 (c++ compilation)
vs2015 installation
OpenCV3.2 configuration
Driver installation, download the corresponding version from NVIDIA's official website
cuda choose 9.0, cudnn package
configure and compile
Double-click \darknet-master\build\darknet\darknet.sln (if there is no GPU, use the darknet_no_gpu.sln project) to open the project
- Modify the words related to the cuda version in darknet.vcxproj
- In the darknet property page, modify the OpenCV section in the c/c++ and additional include directories in the linker
- compile
Second, using Ubuntu 16.04 to achieve
Download and install directly
git clone https://github.com/pjreddie/darknet
cd darknet/
make
Modify the MakeFile file GPU/CUDNN/OPENCV, etc. as needed
3. Both Ubuntu16.04 and Windows can be used
1 hard+tebsorflow-gpu 版
install anaconda3
Install keras and TensorFlow-gpu
pip install tensorflow-gpu
pip install keras
install yad2k
git clone https://github.com/allanzelener/yad2k.git cd yad2k conda env create -f environment.yml source activate yad2k pip install numpy h5py pillow
2、TensorFlow
Compile and install
git clone https://github.com/thtrieu/darkflow
cd darknet/
python3 setup.py build_ext --inplace
pip install .