Kan.py Apr 2026

While more parameter-efficient for some tasks, the current implementation is often slower than optimized MLPs.

(often referred to as pykan ) is the official Python implementation of Kolmogorov-Arnold Networks (KANs) , a novel neural network architecture inspired by the Kolmogorov-Arnold representation theorem. Unlike traditional Multi-Layer Perceptrons (MLPs) that use fixed activation functions on "neurons" (nodes), KANs place learnable activation functions—typically splines—directly on the "weights" (edges) of the network. Core Concept: The KAN Architecture kan.py

: It is designed to mimic the structure of standard PyTorch models, allowing users to define a model with simple parameters like width , grid (spline resolution), and k (spline order). While more parameter-efficient for some tasks, the current

The fundamental shift in KANs is the replacement of fixed linear weights with univariate functions. Core Concept: The KAN Architecture : It is

: In a standard MLP, a connection is just a single number (