# atong01/conditional-flow-matching

TorchCFM: a Conditional Flow Matching library

Repository: https://github.com/atong01/conditional-flow-matching
Canonical: https://ross.abutalabs.com/products/conditional-flow-matching
Homepage: https://arxiv.org/abs/2302.00482
Language: Python
License: MIT
License Family: permissive
Topics: continuous-normalizing-flows, optimal-transport, pytorch
Last push: 2026-07-20T22:02:27+00:00

## Health v2 (maintenance only)
Score: 74/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 93, release rhythm 40, longevity 93
- inputs: {"age_days": 1311, "days_push": 44, "days_rel": 541, "gap_med": 5, "n_releases_24m": 2}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 2571, forks 223 (observed 2026-08-28T04:07:01.343190+00:00)

## What it is
TorchCFM is a PyTorch library implementing Conditional Flow Matching (CFM), a simulation-free training objective for continuous normalizing flow generative models. It includes optimal transport variants (OT-CFM) and examples for image generation, single-cell dynamics, and tabular data.

## Use cases
- train continuous normalizing flows without simulation
- train flow matching generative models in pytorch
- use optimal transport conditional flow matching
- generate images with flow matching instead of diffusion
- model single-cell dynamics with CNFs
- speed up training and inference of normalizing flows

## When to choose
- you want a fast, simulation-free alternative to diffusion model training
- you need deterministic flows with efficient inference
- you want OT-based couplings for straighter, more stable flows
- you work in PyTorch on generative modeling research

## When to avoid
- you need a production image-generation pipeline rather than a research library
- you prefer standard diffusion frameworks with large pretrained model ecosystems
- you don't use PyTorch

## Facets
- artifact type: library
- maturity: active
- function: machine-learning, deep-learning, simulation
- domain: machine-learning, deep-learning, artificial-intelligence, data-science
- platform: python
- tags: flow-matching, continuous-normalizing-flows, optimal-transport, generative-models, pytorch, diffusion-models, generative-modeling, gpu

## Member repositories
- atong01/conditional-flow-matching (main) score 74

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:07:01.343190+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 2026-08-30T02:23:17.495088+00:00, confidence not recorded.
  - readme: https://github.com/atong01/conditional-flow-matching (fetched 2026-08-28T04:07:01.343190+00:00, sha 7fc3094e6f84)
  - homepage: https://arxiv.org/abs/2302.00482 (fetched 2026-08-29T10:05:43.887217+00:00, sha 9672b85f35fa)
  - site_page: https://info.arxiv.org/about/donate.html (fetched 2026-08-29T10:05:43.890447+00:00, sha cca9c3a11c56)
  - site_page: https://info.arxiv.org/about/ourmembers.html (fetched 2026-08-29T10:05:43.894039+00:00, sha 47cbc55ff1de)
  - site_page: https://info.arxiv.org/about (fetched 2026-08-29T10:05:43.895886+00:00, sha a1f16f915a9a)
  - site_page: https://info.arxiv.org/labs/index.html (fetched 2026-08-29T10:05:43.892291+00:00, sha b14a8d05a0ec)
- Data as of 2026-08-30T08:39:29.467469+00:00.