Stability AI has introduced Stable Audio Open Small, a cutting-edge AI model designed for audio generation, which runs efficiently on smartphones. It was developed in partnership with Arm and utilizes royalty-free audio sources for training, avoiding copyright issues common with other models. While it can produce short audio samples quickly, it has limitations, including English-only prompts and a focus on Western music styles. Additionally, its usage model varies for different revenue levels, allowing free use for smaller developers, but imposing fees for larger enterprises.
Stable Audio Open Small is a stereo audio-generating AI model, designed for operational efficiency on smartphones, making audio generation more accessible and offline capable.
Developed in collaboration with Arm, Stable Audio Open Small is positioned as the fastest audio-generating model available, utilizing royalty-free music for training and avoiding copyright issues.
With 341 million parameters, the model can produce short audio samples quickly, though it currently has limitations in generating realistic vocals and diversity in musical styles.
Stable Audio Open Small is free for smaller developers but imposes fees on those with higher revenue, which may restrict its uptake in varied commercial scenarios.
Collection
[
|
...
]