# machine-learning
- p
pi
05/15/2021, 1:59 AMoh ok, weird - u
Uberduck Agent
05/15/2021, 7:03 PMSent voice request! - u
Uberduck Agent
05/15/2021, 7:18 PMSent voice request! - h
Hugh G. Rection
05/15/2021, 7:22 PMI submitted the Private (from the Madagascar movies) voice model, again for the 2nd time, because the 1st time is that he didn't make it. So here's what this model would sound like: - h
Hugh G. Rection
05/15/2021, 7:23 PM"Private here, reporting for duty, sir!" That's what the model says on the synthesis notebook. - u
Uberduck Agent
05/15/2021, 8:01 PMYou are missing the Agreed role! Do so in #842148452464853002 - w
Waltman13 (semi-active)
05/15/2021, 8:42 PMmario really needs a better model. either that or his transcript needs to be adjusted. - p
PUMPKINEATER
05/15/2021, 8:51 PMJeez it's not that bad. - w
Waltman13 (semi-active)
05/15/2021, 9:03 PMI know. I mean voice itself doesn't sound bad, game gallery works fine, but I just have a feeling that the transcribing and audio cutting could maybe use a bit of work. - w
Waltman13 (semi-active)
05/15/2021, 9:07 PMI mean there's gotta be some way to stop mario from generating extra data that goes on for a minute.. - z
zwf
05/15/2021, 9:09 PMI actually made a change the other day that decreases the babbling from certain models. (If you try stimpy you can notice that it babbles less) - n
Needlemouse
05/15/2021, 9:09 PMIt’s just luck - z
zwf
05/15/2021, 9:09 PMI'll make that tweak for mario too. - n
Needlemouse
05/15/2021, 9:10 PMIf you want no extra data try not to use symbols ‘ , “ - z
zwf
05/15/2021, 9:13 PMPlease let me know what models you hear babbling and I'll adjust the parameters for those models. - z
zwf
05/15/2021, 9:13 PMI changed it for a bunch of ones I was aware of but I know that I missed some - p
PUMPKINEATER
05/15/2021, 9:14 PMMarge does it too. - p
PUMPKINEATER
05/15/2021, 9:14 PMIt always does it. - p
PUMPKINEATER
05/15/2021, 9:15 PMI mean it can speak english. - i
IBob012
05/15/2021, 9:25 PMmickey also does it - i
Isaac Rodriguez (Odcub)
05/15/2021, 9:25 PMthis is making my audio quiet idk how but here is an example: https://colab.research.google.com/drive/18IJZZDW1NO7KOslg_WMOCrMeiqz9jOYF?usp=sharing - i
IBob012
05/15/2021, 9:29 PMsounds like something that would happen if you get k80 - i
IBob012
05/15/2021, 9:29 PMalso boomerang announcer - c
ccateni
05/16/2021, 2:23 AMquestion: If a character doesn't say sentences and only does grunts and similar voice clips (huh, grr, etc.), is it possible to turn it into a TTS character? - j
JacobLenstar
05/16/2021, 2:23 AMno - j
JacobLenstar
05/16/2021, 2:24 AMyou need a wide range of syllables and words - c
ccateni
05/16/2021, 2:24 AMgot it, I'll look into the va's work and see if I can figure something out for a future TTS dataset. - j
JacobLenstar
05/16/2021, 2:24 AMand at least every letter in the alphabet - j
JacobLenstar
05/16/2021, 2:24 AMif you don’t have an X or a Z, they’re gonna have trouble with words using those letters