Baran-phys/Tropical-Attention

[NeurIPS 2025] Official code for "Tropical Attention: Neural Algorithmic Reasoning for Combinatorial Algorithms"

35
/ 100
Emerging

This project offers a new way to solve complex computational problems in fields like phylogenetics, cryptography, and particle physics. It takes in structured data representing challenges like finding the shortest paths in a graph or optimally packing items, and outputs precise solutions or classifications. It's for researchers and practitioners in computationally intensive fields who need to tackle hard combinatorial problems more efficiently and accurately.

Use this if you are developing AI models for complex tasks that involve combinatorial optimization, graph theory, or resource allocation, and you need sharper, more robust, and faster solutions than traditional neural networks provide.

Not ideal if your primary goal is general-purpose natural language processing or image recognition, as this tool is specifically designed for neural algorithmic reasoning on structured, combinatorial problems.

Combinatorial Optimization Phylogenetics Cryptography Particle Physics Mathematical Discovery
No Package No Dependents
Maintenance 6 / 25
Adoption 7 / 25
Maturity 15 / 25
Community 7 / 25

How are scores calculated?

Stars

27

Forks

2

Language

Python

License

MIT

Last pushed

Oct 23, 2025

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/transformers/Baran-phys/Tropical-Attention"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.