Skip to content

Latest commit

 

History

History
73 lines (51 loc) · 2.15 KB

INSTALL.md

File metadata and controls

73 lines (51 loc) · 2.15 KB
Raw

Installation

The code was tested on Ubuntu 16.04, with Anaconda Python 3.6 and PyTorch v0.4.1. NVIDIA GPUs are needed for both training and testing. After install Anaconda:

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

    conda create --name CenterNet python=3.6

    And activate the environment.

    conda activate CenterNet

  2. Install pytorch0.4.1:

    conda install pytorch=0.4.1 torchvision -c pytorch

    And disable cudnn batch normalization(Due to this issue).

    # PYTORCH=/path/to/pytorch # usually ~/anaconda3/envs/CenterNet/lib/python3.6/site-packages/
# for pytorch v0.4.0
sed -i "1194s/torch\.backends\.cudnn\.enabled/False/g" ${PYTORCH}/torch/nn/functional.py
# for pytorch v0.4.1
sed -i "1254s/torch\.backends\.cudnn\.enabled/False/g" ${PYTORCH}/torch/nn/functional.py

    For other pytorch version, you can manually open torch/nn/functional.py and find the line with torch.batch_norm and replace the torch.backends.cudnn.enabled with False. We observed slight worse training results without doing so.

  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:

    CenterNet_ROOT=/path/to/clone/CenterNet
git clone https://github.com/CaoWGG/CenterNet-CondInst.git $CenterNet_ROOT

  5. Install the requirements

    pip install -r requirements.txt

  6. Compile deformable convolutional (from DCNv2).

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

  7. [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