# machine-learning
- c
CyberClone
07/01/2021, 11:46 PMfor example - c
CyberClone
07/01/2021, 11:46 PMput the error here - f
free_muffin
07/01/2021, 11:59 PMdoes letting the training run for longer give you a better result? - u
user
07/02/2021, 12:02 AMjust look. - u
user
07/02/2021, 12:03 AMit's born.wav, not born.npy - u
user
07/02/2021, 12:08 AMwhy this happens? - i
IBob012
07/02/2021, 12:09 AMits normal, the mel spectogram code should have created the npy files - u
user
07/02/2021, 12:09 AMthe mel spectogram is on 0% - a
Aniball
07/02/2021, 5:30 AMcrackle - a
Aniball
07/02/2021, 5:34 AM:P - w
Waltman13 (semi-active)
07/02/2021, 6:23 AMit's like a vinyl record - n
Needlemouse
07/02/2021, 1:38 PMJust call it a record voice people will believe it - e
Emoto
07/02/2021, 1:54 PMIt keeps spitting out an error message at the 2nd block of code - a
Awesome Face (GTX 1660 Ti)
07/02/2021, 1:54 PMyeah but we have in #841437191073955920 - e
Emoto
07/02/2021, 1:55 PMI dont see any colab relating to waveglow in the channel - a
Awesome Face (GTX 1660 Ti)
07/02/2021, 1:57 PMbut - a
Awesome Face (GTX 1660 Ti)
07/02/2021, 1:57 PMin simplfied training - a
Awesome Face (GTX 1660 Ti)
07/02/2021, 1:57 PMit is waveglow - a
Awesome Face (GTX 1660 Ti)
07/02/2021, 1:57 PMespecially synthesis - e
Emoto
07/02/2021, 2:00 PMWhere exactly? I feel like im blind cause i cant find it lol. I just find the general tacotron2 training colab, not the one that has waveglow - a
Awesome Face (GTX 1660 Ti)
07/02/2021, 2:01 PMhttps://colab.research.google.com/drive/1WTilMdm9Vf7KE79gzkeeTBigAN6iv3Bg?usp=sharing it is tacotron and waveglow training - a
Awesome Face (GTX 1660 Ti)
07/02/2021, 2:01 PMbut simplified - a
Awesome Face (GTX 1660 Ti)
07/02/2021, 2:01 PMhttps://colab.research.google.com/drive/1WTilMdm9Vf7KE79gzkeeTBigAN6iv3Bg?usp=sharing and this is for synthesis for tacotron and waveglow - e
Emoto
07/02/2021, 2:02 PMOh i see, so they train both technically spwaking - a
Awesome Face (GTX 1660 Ti)
07/02/2021, 2:02 PMehm, not really but yeah. - a
Awesome Face (GTX 1660 Ti)
07/02/2021, 2:02 PMit's same as taking list.txt val and train text - a
Awesome Face (GTX 1660 Ti)
07/02/2021, 2:02 PMand datasets - e
Emoto
07/02/2021, 2:03 PMI was kinda looking for training and improving the waveglow model though, and afaik, the simplefied doc has not the option to train the waveglow model by its own