# machine-learning
- u
(Dawn) Will Draw Fictional Women
10/29/2021, 9:04 PMi give it 6 or so more years before it starts getting MASSIVE coverage on rampant pedophilia - u
user
10/29/2021, 9:09 PMhow though? like basically any other metaverse type thing will probably have the same issue - c
Couch
10/29/2021, 9:09 PMroblox is targeted for childs - u
user
10/29/2021, 9:15 PMwell im a dev on ther - i
IBob012
10/29/2021, 9:15 PMmario works on roblox?? - u
user
10/29/2021, 9:16 PMyes XD - t
The Wiggles and High School D×D
10/29/2021, 9:21 PMSo I discovered this - t
The Wiggles and High School D×D
10/29/2021, 9:21 PM - t
The Wiggles and High School D×D
10/29/2021, 9:21 PMAnd I know discovering more voice cloning options isn't the best for my case - t
The Wiggles and High School D×D
10/29/2021, 9:21 PMBut I need to know if this will be the solution for my problems - h
hecko
10/29/2021, 9:31 PMwhat are your problems exactly - t
The Wiggles and High School D×D
10/29/2021, 9:33 PMWell, based on the work I already attempted to do - t
The Wiggles and High School D×D
10/29/2021, 9:33 PMI can't find a good set of English and Japanese datasets that do the job - t
The Wiggles and High School D×D
10/29/2021, 9:33 PMFor encoding and synthesising - t
The Wiggles and High School D×D
10/29/2021, 9:34 PMIt's all very fickle and I can't pinpoint why I can't get the machine to encode and synthesise it (and then there's vocoding which I have yet to reach) - t
The Wiggles and High School D×D
10/29/2021, 9:34 PMI asked somebody for help in the Python server but that didn't go anywhere - t
The Wiggles and High School D×D
10/29/2021, 9:34 PMI might ask again although it's not exactly easy to get help like that on there - t
The Wiggles and High School D×D
10/29/2021, 9:34 PMOr anywhere - t
The Wiggles and High School D×D
10/29/2021, 9:59 PM...given how I saw it being described by another program in that I'm meant to combine my audio into one folder, that could be the issue As long as I can figure out how to create a metadata file from it anyway - j
Justin
10/30/2021, 5:10 AM - h
hecko
10/30/2021, 4:44 PMwow i can't believe tommyinnit made this - g
Gamma Prime
10/30/2021, 5:16 PMI didn't delete the old dataset, so I'm trying to train Starscream again. - g
Gamma Prime
10/30/2021, 5:16 PMI did clean it up and redo transcriptions. - g
Gamma Prime
10/30/2021, 5:17 PMThat's epoch 25 IIRC. - g
Gamma Prime
10/30/2021, 5:18 PMI'm also working on this one. - g
Gamma Prime
10/30/2021, 5:18 PMIt probably needs a bit more data, which I can get. I still have to avoid intense action scenes because they're too noisy. - h
hecko
10/30/2021, 6:50 PMooh this has some pretrained tacotron models for a few languages https://github.com/Kyubyong/css10 - h
hecko
10/30/2021, 6:50 PMidk if it's tacotron 2 though - c
Couch
10/30/2021, 6:53 PMonly one way to find out - g
Gamma Prime
10/30/2021, 6:55 PMStarscream failed again. There was too little progress even after 80 epochs to keep wasting GPU time training it.