# torchsom: The Reference PyTorch Library for Self-Organizing Maps

arXiv: https://arxiv.org/abs/2510.11147 · HTML: https://mlsanity.com/paper/2510.11147/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 0 of its 37 references that are in the map.
- First posted on arXiv 13 Oct 2025.

## Built on by (most cited first, 1 of 1)

- [Self-Organized Conformal Prediction: Reducing Regional Coverage Gaps with Unsupervised Group Discovery](https://mlsanity.com/paper/2606.29403/) (2026)

## Abstract (arXiv)

This paper introduces torchsom, an open-source Python library that provides a reference implementation of the Self-Organizing Map (SOM) in PyTorch. This package offers three main features: (i) dimensionality reduction, (ii) clustering, and (iii) friendly data visualization. It relies on a PyTorch backend, enabling (i) fast and efficient training of SOMs through GPU acceleration, and (ii) easy and scalable integrations with PyTorch ecosystem. Moreover, torchsom follows the scikit-learn API for ease of use and extensibility. The library is released under the Apache 2.0 license with 90% test coverage, and its source code and documentation are available at https://github.com/michelin/TorchSOM.