Meta launches Omnilingual Automatic Speech Recognition (ASR) artificial intelligence (AI) model, capable of automatically recognising over 1600 languages. So, what should you know about it?
It was built using 7 billion parameters
For your information, it was developed using 7 billion parameters and is offered as open source under the Apache 2.0 license. Of these 1600 languages, 78% can be translated with an error rate below 10%. To train Omnilingual ASR, Meta used 249 languages with high usage resources, 881 languages with medium usage resources, and 546 languages with low usage resources.
According to Meta, Omnilingual ASR can theoretically be expanded to support up to 5,400 world languages. This is a far cry from OpenAI's Whisper language translation model, which only supports 99 major languages.
Moreover, Meta added that Omnilingual ASR also supports 500 new languages that are not supported by any previous automatic language recognition (ASR) system. Omnilingual ASR can be used to automatically identify languages in audio and text, with transcripts being produced simultaneously.
Stay tuned to TechNave.com for more updates.
COMMENTS