Flow Matching Python Packages

pi-zero-pytorch

Implementation of π₀, the robotic foundation model architecture proposed by Physical Intelligence

f5-tts-mlx

Implementation of F5-TTS in MLX

torchebm

🍓 Simulation-free, GPU-first generative modeling in PyTorch ⚡ Composable primitives for scalable, stable training of modern EBMs, diffusion, flow matching, and Schrödinger bridges.

transfusion-pytorch

Pytorch implementation of Transfusion, "Predict the Next Token and Diffuse Images with One Multi-Modal Model", from MetaAI

gaia2-pytorch

Implementation of the world model architecture for self driving out of Wayve

medlatents

Generative modelling of tokenized medical images: autoregressive, MaskGIT, diffusion, flow matching, and Bayesian flow.

e2tts-mlx

Embarrassingly Easy Fully Non-Autoregressive Zero-Shot TTS (E2 TTS) in MLX

gensbi

Generative Models for Simulation-Based Inference in JAX

mlx-e2-tts

Implementation of E2-TTS, "Embarrassingly Easy Fully Non-Autoregressive Zero-Shot TTS", in MLX

gensbi-examples

Examples and benchmarks for the GenSBI library

symupe

SyMuPe: Affective and Controllable Symbolic Music Performance (ACM MM '25, Outstanding Paper Award)

flowmatching-bdt

Implementation of flow matching on tabular data using XGBoost

mintflow

Generation of disentangled microenvironment-induced and intrinsic gene expression vectors from spatial transcriptomics data

mlx-minimax-music3

Pure MLX inference for MiniMax Music 3 on Apple silicon.

mixture-of-flows

This work introduces Flow Matching Mixture of Experts (FM-MoE), a framework that replaces conventional MLP experts with flow matching networks. Each expert learns a continuous transformation through an ordinary differential equation (ODE), enabling more expressive feature mappings while maintaining the computational efficiency of sparse experts

multicom-ligand

Comprehensive ensembling of protein-ligand structure and affinity prediction methods (CASP16)

inigen

Generation of disentangled microenvironment-induced and intrinsic gene expression vectors from spatial transcriptomics data