A high-performance library for path signatures and rough path methods on CPU and GPU.
Installation
Get up and running with pySigLib on Windows, Linux, or macOS — with optional CUDA support.
Conventions
Default behaviours, CPU/GPU conventions, parallelism, and data format expectations.
API Reference
Signatures, log signatures, signature kernels, and more.
PyTorch API
Use pySigLib functions as native PyTorch autograd functions with full gradient support.
JAX API
Use pySigLib functions with JAX transformations — jit, grad, and vmap.
C++ Library
Direct access to the underlying C++/CUDA siglib library.
Citation#
If you found this library useful in your research, please consider citing the paper:
@article{shmelev2025pysiglib,
title={pySigLib-Fast Signature-Based Computations on CPU and GPU},
author={Shmelev, Daniil and Salvi, Cristopher},
journal={arXiv preprint arXiv:2509.10613},
year={2025}
}