# machine-learning
- u
(Dawn) Will Draw Fictional Women
09/15/2021, 4:17 AMhold on - u
(Dawn) Will Draw Fictional Women
09/15/2021, 4:18 AMit only started babbling on this prompt - u
(Dawn) Will Draw Fictional Women
09/15/2021, 4:18 AMall the other ones i tossed at it are working fine - u
(Dawn) Will Draw Fictional Women
09/15/2021, 4:20 AMfuck i see the problem - u
(Dawn) Will Draw Fictional Women
09/15/2021, 4:20 AMthe dataset got mixed up somehow - u
(Dawn) Will Draw Fictional Women
09/15/2021, 4:21 AMthe wrong lines are on the wrong voiceclips - u
(Dawn) Will Draw Fictional Women
09/15/2021, 4:21 AMwelp! time to retrain! - u
(Dawn) Will Draw Fictional Women
09/15/2021, 4:32 AMhopefully i corrected all the errors - u
(Dawn) Will Draw Fictional Women
09/15/2021, 4:33 AMthe model sounded pretty good when it used the clips that were outside the faulty data though! - u
(Dawn) Will Draw Fictional Women
09/15/2021, 4:42 AMyeaah yeaah muuuch better - j
JacobLenstar
09/15/2021, 1:12 PMwhen a model stonks out on the first epoch - u
user
09/15/2021, 8:58 PM - t
The Wiggles and High School D×D
09/15/2021, 11:59 PMI keep trying to train the LJ dataset as per the cross lingual model listed above but I'm stuck It keeps telling me that I'm missing a file even when I added it in - j
JacobLenstar
09/16/2021, 2:17 PMAlvin is not doing well - j
JacobLenstar
09/16/2021, 2:19 PM"ALVIN!!!" - j
JacobLenstar
09/16/2021, 2:19 PMI refuse to train, dave! - j
JacobLenstar
09/16/2021, 2:21 PMthere is no glitz and glamour in the text, to speech market! - j
JacobLenstar
09/16/2021, 2:21 PMthis is gonna be such a fun voice when it trains correctly lol - j
JacobLenstar
09/16/2021, 2:55 PM119 epoches into alvin - u
(Dawn) Will Draw Fictional Women
09/16/2021, 2:58 PMa chipmunk-sized stonk - u
(Dawn) Will Draw Fictional Women
09/16/2021, 3:00 PMperfect - t
The Wiggles and High School D×D
09/16/2021, 6:41 PMI'm stuck on designing the model again This time, I can't train the model on any level because I was missing a file before But it still won't work even after I insert the file in the dataset - c
CliffracerX
09/16/2021, 9:12 PMHopping in here to ask for advice on training - specifically, how to (if at all) deal with this omnipresent buzzing that's been afflicting anything I make w/ TalkNet so far. I have a feeling it's the fault of my datasets being small (a whack at Dagoth Ur from TES:III Morrowind, which gives it all of a minute's worth to learn from, and a whack at the narrator in Receiver's original set of tapes, which gives it about eight and a half minutes), but figured I should ask the experts just in case. - c
CliffracerX
09/16/2021, 9:13 PMA demo of Dagoth Ur - c
CliffracerX
09/16/2021, 9:13 PMand the Receiver tape-reader, which sounds significantly better but still a little buzzy (this may be because the tapes don't have reverb and Dagoth Ur has lots of it, admittedly) - u
(Dawn) Will Draw Fictional Women
09/16/2021, 10:43 PMis this what the children call stocks - c
CliffracerX
09/16/2021, 10:47 PMgiven how touchy training that can be, I do believe a decent diagonal qualifies as Stonks, yes - t
Tiege
09/16/2021, 11:00 PMHey wait, do the Pony Preservation Projects anons not like Uberduck and Vocodes? That sucks - u
(Dawn) Will Draw Fictional Women
09/16/2021, 11:01 PM>makes a damon albarn dataset quickly