Stability AI launches an audio generating model that can work on smartphones

AI startup stability AI has released A stable audio open small, Stereo audio generating AI model that the company claims is the fastest on the market-and is efficient enough to work on smartphones.

The stable audio open small is the fruit of cooperation between AI and ARM for stability, the chipmaker, which produces many processors inside tablets, phones and other mobile devices. While a number of AI-powered applications can generate audio such as Suno and UDIO, most rely on cloud processing, which means that they cannot be used offline.

Stability also claims that the stable Audio Open Small training kit is made up entirely of songs from the free audio libraries of Audio and Freesound. This is unlike the training sets of the aforementioned Sun and Udio, which according to the messages contain Copyright contentPutting IP risk.

The stable Audio Open Small is 341 million parameters and optimized to perform ARM processors. (The parameters, sometimes called “weights”, are the internal components of a model that directs its behavior.) Separate for the rapid generation of short audio samples and sound effects (eg drum and tools), the stable audio open small can produce up to 11 seconds an audio for less than 8 seconds, hard stability of AI.

Here is an excerpt generated by a stable audio open small:

And here’s one more:

The model is not without its limitations. The stable audio open Small keeps prompted prompts written in English and stability notes in its documentation that the model cannot generate realistic vocals or high quality songs. The model also does not perform equally well in the music styles, warns stability-the result of its data on training, addicted to the West.

In another potential wrinkle for Devs, the stable Audio Open Small has somewhat restrictive conditions for use. It is free to use for researchers, amateurs and businesses with less than $ 1 million annual revenue, but developers and organizations that make revenue of over $ 1 million have to pay for stability enterprise licenseS

Stability, the dressed company behind the popular diffusion of the image generation model, raised new money last year As investors, including Eric Schmidt and Napster Founder Sean Parker, tried to turn the business. Emad Mostaque co -founder, Stability stability and former CEO, announces that the financial devastation stability has been managed, which causes staff to resign, partnership with Canva, to collapse investors to bother with the company’s prospects.

In the last few months, stability has hired a new CEO, appointed director James Cameron to his board of directors and has launched several new models to generate images.