# 🎃general🎃
- u
user
02/03/2021, 9:07 PMWe need more i think someone didnt joined - u
user
02/03/2021, 9:08 PM - u
user
02/03/2021, 9:09 PMPeople choose to join at their own leisure. - u
user
02/03/2021, 11:57 PM - z
zwf
02/04/2021, 3:44 AMrelevant - u
user
02/04/2021, 3:52 AMIs that Michael Rosen?? Did you made it ?? - z
zwf
02/04/2021, 3:58 AMthat is michael rosen, but it's the ground truth, not synthetic - u
user
02/04/2021, 3:59 AMDid you download the Michael Rosen data? - z
zwf
02/04/2021, 3:59 AMyea I did, I'm working on the model now - u
user
02/04/2021, 3:59 AMOk - u
user
02/04/2021, 4:00 AMAre you gonna add it on your web site? - z
zwf
02/04/2021, 4:00 AMyeah, as soon as the quality is decent - z
zwf
02/04/2021, 4:00 AMI'll keep you posted - u
user
02/04/2021, 4:01 AMVocodes Michael Rosen sounds not very good - u
user
02/04/2021, 4:01 AMHe can't talk - u
user
02/04/2021, 4:03 AMSounds very weird - e
echelon
02/04/2021, 4:04 AMIt's glow-tts - u
user
02/04/2021, 4:04 AMI made a new data - e
echelon
02/04/2021, 4:05 AMTacotron and other models sound much smoother - u
user
02/04/2021, 4:06 AMWhat happens if you clone a music sound - e
echelon
02/04/2021, 4:06 AMBut I didn't want to pay for GPUs. Unfortunately glow-tts isn't very amenable to multi speaker models, so I wound up paying for RAM - u
user
02/04/2021, 4:06 AMWhat GPU you use? - e
echelon
02/04/2021, 4:06 AMI don't use GPUs. Glow is slim enough to do real time inference on CPU - u
user
02/04/2021, 4:07 AMCPU? - e
echelon
02/04/2021, 4:07 AMBut it doesn't sound the greatest. It's a trade-off - u
user
02/04/2021, 4:07 AMBut your Dr Phil sounds good - e
echelon
02/04/2021, 4:08 AMDifferent voices train better than others 😕 - u
user
02/04/2021, 4:13 AMI made the Michael Stevens I don't mean Michael Rosen. But I used a video editor on my phone and worked the data - g
Gosmokeless28
02/04/2021, 9:57 PMHi - j
Justin
02/07/2021, 7:20 PMJust gotta transfer this onto my body..see the results..eh?