AI Startup Stability AI has issued Stable Audio Open Small, a “stereo” audio-generating AI model that the company claims the fastest on the market is and efficient enough to run on smartphones.
Stable Audio Open Small is the fruit of a collaboration between stability AI and Arm, the chip maker who produces many of the processors in tablets, telephones and other mobile devices. Although a number of AI-driven apps can generate audio, such as Suno and Udio, they are most dependent on cloud processing, which means that they cannot be used offline.
Stability also claims that the Stable Audio Open Small training set consists entirely of songs from the Royalty-Free Audio Libraries Free Music Archive and Freesound. In contrast to the training sets of the aforementioned Suno and Udio, which is said to contain copyrighted content, which means that an IP risk entails.
Stable Audio Open Small is 341 million parameters in size and optimized to be performed on Arm CPUs. (Parameters, sometimes referred to as weights, are the internal components of a model that accompanies its behavior.) For the rapid generation of short audio monsters and sound effects (eg drum and instruments), stable audio opened Small can up to 11 seconds audio in less than 8 seconds.
Here a monster has been generated by stable audio open small:
And here there is another:
The model is not without limitations. Stable Audio Open Small only supports instructions written in English, and stability notes in its documentation That the model cannot generate realistic singing or songs of high quality. The model also does not perform well about music styles, stability warns a consequence of the Western training data.
In another potential wrinkle for developers, stable Audio Open Small has somewhat restricting conditions of use. It is free to use for researchers, hobbyists and companies with less than $ 1 million in annual turnover, but developers and organizations that earn more than $ 1 million in income have to pay for stability Enterprise License.
Stability, the besieged company behind the popular image generation model stable diffusion, increased new money as investors last year, including Eric Schmidt and founder of Napster Sean Parker, wanted to turn the company. Emad Mostaque, co-founder of stability and ex-CEO, reportedly incorrectly managed stability in financial ruin, as a result of which the staff resigned, a partnership with Canva to fall through and investors are concerned about the prospects of the company.
In recent months, stability has hired a new CEO, Titanic Director James Cameron collected to his board of directors and released various new models for generating image generation.
#Stability #releases #audio #fence #model #performed #smartphones #Techcrunch
