Skip to content

Latest commit

 

History

History
76 lines (53 loc) · 1.78 KB

INSTALL.md

File metadata and controls

76 lines (53 loc) · 1.78 KB
Raw

Installation

The code was tested on Ubuntu 16.04, with Anaconda Python 3.7 and PyTorch v1.4.0. NVIDIA GPUs are needed for both training and testing. After install Anaconda:

  1. [Optional but recommended] create a new conda environment.

    conda create --name CHPDet python=3.7

    And activate the environment.

    conda activate CHPDet

  2. Install pytorch 1.4.0:

  3. Install COCOAPI:

    # COCOAPI=/path/to/clone/cocoapi
git clone https://github.com/cocodataset/cocoapi.git $COCOAPI
cd $COCOAPI/PythonAPI
make
python setup.py install --user

  4. Clone this repo:

    CHPDet_ROOT=/path/to/clone/CHPDet
git clone https://github.com/zf020114/CHPDet $CHPDet_ROOT

  5. Install the requirements

    pip install -r requirements.txt

  6. Compile deformable convolutional (from DCNv2).

    cd $CHPDet_ROOT/src/lib/models/networks/DCNv2
./make.sh

  7. compile orn from s2anet cd $CHPDet_ROOT/src/lib/models/networks/orn bash make.sh

7.install DOTA_devkit to get the evel result 7.1. install swig

    sudo apt-get install swig

7.2. create the c++ extension for python

    swig -c++ -python polyiou.i
    python setup.py build_ext --inplace

  1. [Optional, only required if you are using extremenet or multi-scale testing] Compile NMS if your want to use multi-scale testing or test ExtremeNet.

    cd $CenterNet_ROOT/src/lib/external
make

  2. Download pertained models for detection or pose estimation and move them to $CHPDet_ROOT/models/.