# hkchengrex/MMAudio

[CVPR 2025] MMAudio: Taming Multimodal Joint Training for High-Quality Video-to-Audio Synthesis

Repository: https://github.com/hkchengrex/MMAudio
Canonical: https://ross.abutalabs.com/products/mmaudio
Homepage: https://hkchengrex.com/MMAudio/
Language: Python
License: MIT
License Family: permissive
Topics: audio, audio-synthesis, computer-vision, deep-learning, video-to-audio, text-to-audio
Last push: 2026-02-23T06:09:19+00:00

## Health v2 (maintenance only)
Score: 43/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 69, release rhythm 8, longevity 45
- inputs: {"age_days": 634, "days_push": 191, "days_rel": 634, "gap_med": null, "n_releases_24m": 1}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 2264, forks 266 (observed 2026-08-28T04:06:31.988319+00:00)

## What it is
MMAudio is a PyTorch-based model for generating synchronized audio from video and/or text inputs, using multimodal joint training across audio-visual and audio-text datasets. It includes a synchronization module that aligns generated audio with video frames, published as a CVPR 2025 paper.

## Use cases
- generate sound effects for silent videos
- add synchronized audio to AI-generated video from Sora or Veo
- synthesize audio from a text prompt
- dub foley sounds onto video footage automatically
- research video-to-audio synthesis models

## When to choose
- you need to generate audio tracks synced to video content
- you want state-of-the-art video-to-audio or text-to-audio synthesis
- you have a CUDA GPU and want to run inference locally or via demo services

## When to avoid
- you need real-time audio generation on CPU-only hardware
- you need production audio editing rather than generative synthesis
- you require Windows or macOS support, since only Ubuntu is tested

## Facets
- artifact type: library
- maturity: active
- function: audio-processing, machine-learning, deep-learning, video-processing
- domain: artificial-intelligence, deep-learning, machine-learning
- platform: python
- tags: video-to-audio, text-to-audio, audio-synthesis, generative-ai, cvpr-2025, pytorch, audio, video, linux, gpu

## Member repositories
- hkchengrex/MMAudio (main) score 43

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:06:31.988319+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 2026-08-30T02:42:26.847548+00:00, confidence not recorded.
  - readme: https://github.com/hkchengrex/MMAudio (fetched 2026-08-28T04:06:31.988319+00:00, sha 3cbee1cc9efb)
  - homepage: https://hkchengrex.com/MMAudio/ (fetched 2026-08-29T10:22:47.114672+00:00, sha 9cbfb880e66e)
- Data as of 2026-08-30T08:39:29.467469+00:00.