# machine-learning
- b
Balloonya from Pokémon
05/25/2021, 4:37 PMI'm reducing the noise. - p
pi
05/25/2021, 4:38 PMcan I see the dataset you used for that - b
Balloonya from Pokémon
05/25/2021, 4:41 PMNo, because i deleted it. - b
Balloonya from Pokémon
05/25/2021, 4:41 PMI gonna try again - p
pi
05/25/2021, 4:41 PMargh wtf - p
pi
05/25/2021, 4:41 PMI wanted to try a dc tts - p
pi
05/25/2021, 4:42 PM(another tts architecture for which I found a portugese model) - p
pi
05/25/2021, 4:42 PMcan't you restore it? - b
Balloonya from Pokémon
05/25/2021, 4:42 PMi'm gonna recreate it. - b
Balloonya from Pokémon
05/25/2021, 5:21 PMI restored my dataset. - p
pi
05/25/2021, 5:21 PMi'll try it out - r
RGM - ERIOL V2 MODEL ON DEV
05/25/2021, 5:29 PMso, guys, i made an brazilian ai model with lu9, but i don't know if the guys at uberduck will accept this thing this are the results - i
IBob012
05/25/2021, 5:30 PMthis submission is HACKED only brazilians can judge it 😎 whats being said in the audio files - r
RGM - ERIOL V2 MODEL ON DEV
05/25/2021, 5:30 PMi used 39 voice clips from the original brazilian dub - r
RGM - ERIOL V2 MODEL ON DEV
05/25/2021, 5:30 PMi can send you the voice clips in zip on dm - r
RGM - ERIOL V2 MODEL ON DEV
05/25/2021, 5:32 PMi was originally going to submit via the submit model thing from the website - i
IBob012
05/25/2021, 5:32 PMnah i meant whats being said in the result audio - r
RGM - ERIOL V2 MODEL ON DEV
05/25/2021, 5:32 PMdon't remember, oops - r
RGM - ERIOL V2 MODEL ON DEV
05/25/2021, 5:32 PMlet me try again - r
RGM - ERIOL V2 MODEL ON DEV
05/25/2021, 5:33 PM"Todo mundo sabe como eh um bom dia." - r
RGM - ERIOL V2 MODEL ON DEV
05/25/2021, 5:33 PM"Eh um dia bem divertido pra estar hoje." - i
IBob012
05/25/2021, 5:34 PMi think the voice is there but its having a hard time saying anything - r
RGM - ERIOL V2 MODEL ON DEV
05/25/2021, 5:34 PM"Eh que eu escutei uma musica super legal." - r
RGM - ERIOL V2 MODEL ON DEV
05/25/2021, 5:35 PMyou are right, that's bizarre - i
IBob012
05/25/2021, 5:36 PMafter seeing lu9's patolino model i would make 140 the minimum amount of audio files - r
RGM - ERIOL V2 MODEL ON DEV
05/25/2021, 5:37 PMok, i have another episode of the series with more clips, so i can use that as well - i
IBob012
05/25/2021, 5:37 PMfun fact i was trying to do brazilian goku (wendel obviously) but i gave up since his clips always had remaining sound effects - p
pi
05/25/2021, 5:38 PMdid you share the dataset? - p
pi
05/25/2021, 5:39 PMI'd like to see how you substituted stuff - i
IBob012
05/25/2021, 5:39 PMhes getting help from lu9 so i think its similiar to how he did it