# facebookresearch/flow_matching

A PyTorch library for implementing flow matching algorithms, featuring continuous and discrete flow matching implementations. It includes practical examples for both text and image modalities.

Repository: https://github.com/facebookresearch/flow_matching
Canonical: https://ross.abutalabs.com/products/flow_matching
Homepage: http://facebookresearch.github.io/flow_matching
Language: Python
License: NOASSERTION
License Family: other
Last push: 2026-01-05T21:56:33+00:00

## Health v2 (maintenance only)
Score: 48/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 60, release rhythm 35, longevity 45
- inputs: {"age_days": 634, "days_push": 240, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 4705, forks 369 (observed 2026-08-28T04:08:57.254880+00:00)

## What it is
A PyTorch library for implementing flow matching algorithms with continuous, discrete, and Riemannian flow matching implementations. It accompanies the Flow Matching Guide and Codebase and includes training examples for text and image modalities.

## Use cases
- train a flow matching generative model
- implement continuous normalizing flows in pytorch
- discrete flow matching for text generation
- learn flow matching from a guide and codebase
- train a generative image model on CIFAR10
- riemannian flow matching on manifolds

## When to choose
- you want a maintained, research-backed PyTorch implementation of flow matching
- you need both continuous and discrete flow matching with solvers, paths, and schedulers
- you want runnable training examples for text and image modalities

## When to avoid
- you need pre-trained models - none are released
- you need a permissively licensed library for commercial use - it is CC BY-NC 4.0
- you use diffusion models or other generative approaches instead of flow matching

## Facets
- artifact type: library
- maturity: active
- function: machine-learning, deep-learning
- domain: machine-learning, deep-learning, artificial-intelligence
- platform: python
- tags: flow-matching, generative-modeling, pytorch, generative-ai, research

## Member repositories
- facebookresearch/flow_matching (main) score 48

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:08:57.254880+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T18:19:11.100317+00:00, confidence not recorded.
  - readme: https://github.com/facebookresearch/flow_matching (fetched 2026-08-28T04:08:57.254880+00:00, sha f48531befdde)
  - homepage: http://facebookresearch.github.io/flow_matching (fetched 2026-08-29T09:03:04.027843+00:00, sha ae7eb854e3a3)
  - registry_pypi: https://pypi.org/pypi/flow_matching/json (fetched 2026-08-29T09:03:04.037515+00:00, sha 2964085e5645)
- Data as of 2026-08-30T08:39:29.467469+00:00.