flownet2-pytorch-module

Pytorch implementation of FlowNet 2.0: Evolution of Optical Flow Estimation with Deep Networks.

Rearrange all code of nvidia flownet2-pytorch repository for simplicity.

Let flownet be a module that can be plugged into any code easily.

Compile and install

• Activate your python envs.

  Compile testing successfully on the env settings:

  linux centos-7,
  gcc 5.5,
  python 3.6,
  pytorch-1.6,
  cuda-10.2

  You may need to install some required packages(e.g. python base packages, pytorch) and configure env(e.g. cuda version matches gcc version and pytorch version) correctly.

• Compile src code then install on the activated env with running the following bash cmd.

  sh tools/install.sh

Usage

• Download FlowNet2 pretrained weight.
• Test the installation and compile with following cmd:

  python tools/flownet_test.py 

• You can plug it into your code easily.

Related repository

• https://github.com/NVIDIA/flownet2-pytorch