# machine-learning
- e
Emoto
02/09/2022, 9:15 AMSo found the one for fastspeech but not fastpitch, so I doubt they actually have it - h
haru0l
02/09/2022, 2:55 PMoh maybe i misremembered it orz - t
The Wiggles and High School D×D
02/09/2022, 3:26 PM - t
The Wiggles and High School D×D
02/09/2022, 3:26 PMFor my case, would this be a better choice to use? - t
The Wiggles and High School D×D
02/09/2022, 3:26 PMIt's overwhelming in scope and I don't know where to begin - t
The Wiggles and High School D×D
02/09/2022, 3:27 PMBut could this be an easier approach to making my bilingual voice model? - e
Emoto
02/09/2022, 4:29 PMI havent seen a way to train own voices with coqui yet - t
The Wiggles and High School D×D
02/09/2022, 4:36 PMOh - m
mega b
02/09/2022, 4:47 PMI trained some coqui models, as of right now, it might be the best choice to train some multispeaker models - t
The Wiggles and High School D×D
02/09/2022, 4:53 PMDoesn't coqui do multi-speaker? Or is that your point, to use the multispeaker option? - t
The Wiggles and High School D×D
02/09/2022, 7:14 PMReading through to see if Colab is an option - l
Liam Taheny
02/09/2022, 7:16 PMHow do I load WaveGlow with Fast Pitch? - e
Emoto
02/09/2022, 9:01 PMlike it is possible, but i saw no colab nor documentation on how to train a new couqi model or how to set it up for own models - m
mega b
02/09/2022, 9:03 PM - t
The Wiggles and High School D×D
02/09/2022, 9:04 PMI saw pynb files in the repository but they seem to work in individual projects of some kind - m
mega b
02/09/2022, 9:04 PMCheck the link I sent - g
gen
02/10/2022, 12:26 AMIs that like Real Time Voice Cloning? - l
LindosoLord
02/10/2022, 12:29 AMhelp on this? - j
Justin
02/10/2022, 4:28 PMrestart the runtime and try again - t
The Wiggles and High School D×D
02/10/2022, 11:40 PMIs it possible to create two tts models and somehow merge them to be about two different languages? Sorry, I'm just trying to figure out how to use whatever potential multilingual process there is. I see it's a relatively new implementation. - g
Gosmokeless28
02/10/2022, 11:41 PMProbably not. - t
The Wiggles and High School D×D
02/10/2022, 11:41 PMI did notice my question makes no actual sense - t
The Wiggles and High School D×D
02/10/2022, 11:41 PMThere wouldn't be any trained overlap between the two languages - t
The Wiggles and High School D×D
02/10/2022, 11:42 PMBut I don't see anything on how to train with both languages in mind on two metadata files for two speakers Unless I could turn it into one file of two languages...? - t
The Wiggles and High School D×D
02/11/2022, 3:41 PMCopy codetext_cleaner="phoneme_cleaners", use_phonemes=True, phoneme_language="en-us" - t
The Wiggles and High School D×D
02/11/2022, 3:41 PMSo I'm looking at the code for the multi-speaker portion that was given in the template (thanks mega b) - t
The Wiggles and High School D×D
02/11/2022, 3:42 PMThis is the section I can't wrap my head around - t
The Wiggles and High School D×D
02/11/2022, 3:42 PMIf I want to get both English and Japanese as a result - t
The Wiggles and High School D×D
02/11/2022, 3:43 PMI will say that I still have my g2p file that registers an ARPAbet interpretation of English and Japanese - t
The Wiggles and High School D×D
02/11/2022, 3:43 PMI wonder if I could apply that here and have the language selection not be an issue once I combine my datasets