Optimal Transport Python Packages

pot

POT : Python Optimal Transport

point-cloud-utils

An easy-to-use Python library for processing and manipulating 3D point clouds and meshes.

ott-jax

Optimal transport tools implemented with the JAX framework, to solve large scale matching problems of any flavor.

torchcfm

TorchCFM: a Conditional Flow Matching library

torchdr

TorchDR - PyTorch Dimensionality Reduction

squarenet

Sparse local operations for point clouds in any dimension.

moscot

Multi-omic single-cell optimal transport tools

mosaicraft

Perceptual photomosaic generator — Oklab color matching, MKL optimal transport, Hungarian placement, and Oklch tile-pool expansion + recoloring

wot

A software package for analyzing snapshots of developmental processes

torchgw

Fast Sampled Gromov-Wasserstein optimal transport solver — pure PyTorch, scalable, differentiable

pyapd

Computing (optimal) anisotropic power diagrams using GPU acceleration

ggml-ot

Supervised OT with Global Ground Metric Learning

transmorph

A unifying data integration framework.

torchcdt

A differentiable implementation of the (Radon-) Cumulative Distribution Transform in PyTorch

fugw

Scalable python GPU solvers for fused unbalanced gromov-wasserstein optimal transport problems, with routines and examples to align brain data (fMRI)

ot1d

OT1D: Discrete Optimal Transport in 1D by Linear Programming

flash-sinkhorn

Sinkhorn optimal transport kernels in PyTorch + Triton (squared Euclidean, no cost matrix materialization).

jgot

Dynamic optimal transport on sparse reversible graphs in JAX

cmonge

Learning conditional optimal transport maps

wtk

A Wasserstein Subsequence Kernel for Time Series.

wsingular

Python package for the ICML 2022 paper "Unsupervised Ground Metric Learning Using Wasserstein Singular Vectors".

spatrack

An optimal transport-based algorithm for inferring cell trajectories of spatial transcriptomics.

optiflow

Reconstruction of flows from kymographs with optimal transport

insurance-fairness-ot

Discrimination-free insurance pricing via Lindholm marginalisation, causal path decomposition, and Wasserstein barycenter correction