Deep graph matching meets mixed-integer linear programming: Relax at your own risk ?
This repository contains implementation of the paper: Deep graph matching meets mixed-integer linear programming: Relax at your own risk ?
The project directory is composed as follows:
experiments
: contains configuration files to reproduce the results reported in the paper;models
: contains the implementation of the model based on difference graph matching solvers, such as DIP(ours), power_iteration and sinkhornutils
: several utils;
Get started
Preliminary
- Check if
findutils
(>=4.7.0) is available - Check if hdf5 is installed (apt install libhdf5-serial-dev)
- Check if cuda 10.1 and cudnn 7 are available
- Check if texlive-latex-extra is installed (
apt install texlive-latex-extra
) - Check if
torch_geometric
is installed - Check if
gurobipy
is installed
Download data sets
Run the following command to download data sets PascalVOC and SPair-71K
chmod +x ./download_data.sh && ./download_data.sh
Training
Run training and evaluation
python3 train_eval.py path/to/your/json
where path/to/your/json
is the path to your configuration files.
Our experimental configuration files are in ./experiments
.