3 ms·
Thanks! Do you think they're using models as good as the ones echelon uses at FakeYou? Maybe he can get by with less data.
by kragen 5y ago
Thanks! Do you think they're using models as good as the ones echelon uses at FakeYou? Maybe he can get by with less data.
- echelon 5y agoThe best sounding limited data models have at least thirty minutes of audio data and have a similar pitch and timbre to a base data set. You can get by with less, but it requires finesse.