Etika

need to change the dataset a little bit, too much yelling audio

why tf is it so staticy

lemme send

dataset: https://mega.nz/file/2BgzUAIZ#hh_yw4CZji036eCtCFO9W-CXSDZG0tOxRHVPita15Ww if anyone wants to fix it send me the final model

i think you're right

yeah good idea

eh i'll do it tomorrow

Joycon Boyz!

rip.

he really had to do it to himself

Defo, I need to clean up the samples tho it would sound much better

the audio sounds a little clipped but its *much* better than mine

i'll swap the link and credit you if you want

yeah i've noticed the thing tends to take an average of all the samples when I feel there should be specified vocal moods (yelling, neutral, whispering, etc) that the inferencer should be looking at

i wish dev was in chat i would love to talk to them

rip

aight, i'll also see if more training would iron out the issues but i'm not sure

kk

Rip