Discretize speech to unit

#9
by shuaishuaiye - opened

Hello, I would like to use the mhubert model to generate speech units., and may I ask if this "mhubert147_faiss.index" matches the features of the 9th layer in the models of the iter2?

UTTER - Unified Transcription and Translation for Extended Reality org

Hello,

Thanks for the interest on our model! :)
Yes, mhubert147_faiss.index corresponds to the features extracted from the 9th layer of the second iteration.

Thank you very much !

mzboito changed discussion status to closed

Sign up or log in to comment