5 ms·
It's very related to LLMs. Though instead of text tokens you are working with audio tokens (e.g. from SoundStream). Then you go to audio corpus, instead of text
by piecerough 2y ago
It's very related to LLMs. Though instead of text tokens you are working with audio tokens (e.g. from SoundStream). Then you go to audio corpus, instead of text corpus.