INSTALL.md

Installation

Slighty modified from the CenterNet installation guide. Install the DCNv2_latest externally instead of within the repository. Tested on Ubuntu 20.04 with Python3.8 and PyTorch 1.13.1. NVIDIA GPU(s) is(are) needed for both training and testing.

After install Anaconda:

0. [Optional but recommended] create a new environment.

   conda create -n kgn python=3.8
   

   And activate the environment.

   conda activate kgn
   

1. Install pytorch. For pytorch 1.13.1:

   conda install pytorch==1.13.1 torchvision==0.14.1 torchaudio==0.13.1 pytorch-cuda=11.7 -c pytorch -c nvidia
   

   Manually open torch/nn/functional.py and find the line with torch.batch_norm and replace the torch.backends.cudnn.enabled with False. ??

   For the pytorch v0.4.0 and v0.4.1, 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
   

2. Clone this repo:

   KGN_ROOT=/path/to/clone/kgn
   git clone https://github.com/ $KGN_ROOT
   

3. Install the requirements

   pip install -r requirements.txt
   

4. Install and Compile Deformable Convolutional Networks V2 from DCNv2_latest.

   # DCNv2=/path/to/clone/DCNv2
   git clone https://github.com/jinfagang/DCNv2_latest $DCNv2
   cd $DCNv2
   python setup.py install --user
   

5. [Optional, only required if you are using extremenet or multi-scale testing] Compile NMS.

   cd ./src/lib/external
   make