• eveninghere@beehaw.org
    link
    fedilink
    arrow-up
    2
    ·
    5 months ago

    If you have pre-trained model or a classical voice matching algorithm as the basis, few samples might suffice.