Thinklab-SJTU/ThinkMatch

A research protocol for deep graph matching.

What it solves

ThinkMatch is a research-oriented framework designed to simplify the development, implementation, and benchmarking of deep graph matching algorithms. It addresses the NP-hard combinatorial problem of finding node-to-node correspondences between graphs, which is essential for tasks like image keypoint matching, molecule matching, and bridging movies with their synopses.

How it works

The library provides a collection of PyTorch-based implementations of state-of-the-art deep graph matching methods (such as GMN, NGM, BBGM, and COMMON). It offers standardized modules and datasets (including PascalVOC-Keypoint and SPair-71k) to allow researchers to develop new algorithms and compare them fairly against existing benchmarks under consistent experimental settings.

Who it’s for

It is primarily intended for researchers and developers who want to develop new graph matching algorithms, publish academic papers, or dive deep into the network details and hyperparameters of existing models.

Highlights

  • Comprehensive Model Library: Includes a wide array of deep graph matching methods from 2018 to 2026.
  • Standardized Benchmarking: Provides detailed performance metrics across multiple datasets like PascalVOC and Willow Object Class.
  • Flexible Matching Settings: Supports Two-Graph Matching (2GM), Multi-Graph Matching (MGM), and Multi-Graph Matching with a Mixture of Modes (MGM3).
  • Docker Support: Offers pre-configured Docker images to bypass complex environment setup requirements.

Related

  • Project
    stanford-star/relbench

    RelBench is an open benchmark for predictive machine learning on relational databases, providing a standardized way to load and evaluate models on multi-table data.

  • Project
    pykt-team/pykt-toolkit

    A PyTorch-based library for training and benchmarking deep learning knowledge tracing models with standardized data preprocessing and multiple model implementations.

  • Project
    iMoonLab/DeepHypergraph

    A PyTorch-based deep learning library for Graph and Hypergraph Neural Networks that supports high-order message passing and Auto-ML tuning.

  • Project
    TorchSpatiotemporal/tsl

    A PyTorch-based library for neural spatiotemporal forecasting that provides a unified framework for data preprocessing, model prototyping, and Graph Neural Network implementation.

  • Project
    openclimatefix/graph_weather

    A PyTorch library implementing various graph neural networks for weather forecasting and data assimilation, including models like GenCast and Aurora.