declare-lab / tangoflux

Super Fast and Faithful Text to Audio Generation with Flow Matching and Clap-Ranked Preference Optimization

  • Public
  • 1.1K runs
  • GitHub
  • Weights
  • Paper
  • License
Run with an API
Playground API Examples README Versions

Run time and cost

This model costs approximately $0.0035 to run on Replicate, or 285 runs per $1, but this varies depending on your inputs. It is also open source and you can run it on your own computer with Docker.

This model runs on Nvidia L40S GPU hardware. Predictions typically complete within 4 seconds.

Readme

TangoFlux: Super Fast and Faithful Text to Audio Generation with Flow Matching and Clap-Ranked Preference Optimization

TangoFlux

Overall Pipeline

TangoFlux consists of FluxTransformer blocks, which are Diffusion Transformers (DiT) and Multimodal Diffusion Transformers (MMDiT) conditioned on a textual prompt and a duration embedding to generate a 44.1kHz audio up to 30 seconds long. TangoFlux learns a rectified flow trajectory to an audio latent representation encoded by a variational autoencoder (VAE). TangoFlux training pipeline consists of three stages: pre-training, fine-tuning, and preference optimization with CRPO. CRPO, particularly, iteratively generates new synthetic data and constructs preference pairs for preference optimization using DPO loss for flow matching.

cover-photo

Citation


@article{Hung2025TangoFlux,
  title = {TangoFlux: Super Fast and Faithful Text to Audio Generation with Flow Matching and Clap-Ranked Preference Optimization},
  author = {Chia-Yu Hung and Navonil Majumder and Zhifeng Kong and Ambuj Mehrish and Rafael Valle and Bryan Catanzaro and Soujanya Poria},
  year = {2025},
  url = {https://openreview.net/attachment?id=tpJPlFTyxd&name=pdf},
  note = {Available at OpenReview}
}