i will try again

it will take 2 mins to start the voice cloning

amd users uses colab

michael stevens

still there?

this one is much better bec i boosted the audio a bit

ok interesting ... yeah the quality is ok but it doesn't resemble him that closely

what I'm hoping will work is basically: - use a technique similar to the real time voice cloning repo - fine-tune the model for a bit longer on a small amount of data

This only needs 10 or 5 or 3 secs to clone a voice

If you clone 35 mins sounds horrible

@zwf does 15 minutes data work on tacotron 2???

I've never tried with that little. I'd be surprised if it worked but I could be wrong

I making another michael stevens data

Hi

I made a voice

Without numbers

ooh nice

share?

I uploaf

D

Does mp3 work?

yeah, wav or aac is better but mp3 can work

Ok

Downloading?

yeah

I worked alot of hours

I removed musics and claps

nice. yeah doing a quick pass it sounds pretty good

Got it