minisom and kohonen-maps
These are competitors: MiniSom is a mature, widely-adopted SOM implementation suitable for production use, while Kohonen-maps is a lesser-known alternative that extends the basic SOM concept with GSOM (Growing Self-Organizing Maps) but lacks practical adoption.
About minisom
JustGlowing/minisom
:red_circle: MiniSom is a minimalistic implementation of the Self Organizing Maps
Built on NumPy with optional Numba JIT acceleration, it enables dimensionality reduction of high-dimensional data into low-dimensional grid visualizations while supporting both online and batch training modes. Features include configurable grid topologies (rectangular/hexagonal), winner neuron detection, model persistence via pickle, and applications spanning clustering, color quantization, and outlier detection. The vectorized NumPy-first design keeps dependencies minimal while maintaining extensibility for researchers and educators.
About kohonen-maps
abhinavralhan/kohonen-maps
Implementation of SOM and GSOM
Related comparisons
Scores updated daily from GitHub, PyPI, and npm data. How scores work