ZeyueT/AudioX

Health v2 · maintenance only

Flags: no_releases no_license

Adoption not part of the score

1552 stars · 142 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

AudioX is a unified multimodal framework for anything-to-audio generation, producing audio and music conditioned on text, video, image, or audio inputs via a Multimodal Adaptive Fusion module. It is the official research code release for an ICLR 2026 paper, trained on the large-scale IF-caps dataset of over 7 million samples.

Use cases

• generate sound effects from a text prompt
• create music from a text description
• generate audio for a silent video
• complete or inpaint missing audio segments
• research multimodal-conditioned audio generation models
• benchmark text-to-audio and text-to-music generation

When to choose

• you need a single model that handles text-to-audio, text-to-music, video-to-audio, and audio inpainting
• you want a state-of-the-art research model with pretrained checkpoints on Hugging Face
• you have a CUDA-capable GPU and want to experiment with multimodal audio generation

When to avoid

• you need production-grade audio synthesis with commercial licensing guarantees (license is not standard)
• you have no GPU and need fast, lightweight audio generation
• you need real-time or low-latency audio synthesis

Facets

library · maturity active

machine-learning deep-learning audio-processing llm-inference artificial-intelligence machine-learning deep-learning python audio-generation text-to-audio text-to-music video-to-audio diffusion multimodal iclr-2026 research audio gpu linux

Member repositories

For agents

markdown · JSON · MCP: product_card(name="ZeyueT/AudioX")