# machine-learning
- u
(Dawn) Will Draw Fictional Women
03/12/2022, 3:30 PMyou can continue training off a checkpoint - u
(Dawn) Will Draw Fictional Women
03/12/2022, 3:30 PMjust make sure the parameters and output destination are the same - j
JamesVincentWalsh
03/12/2022, 3:32 PMI was wondering if it was possible to "compound" the training, but didn't want to get my hopes up. Looks like it is. How do I continue from a checkpoint? - u
(Dawn) Will Draw Fictional Women
03/12/2022, 3:32 PMjust do the whole process again as usual - u
(Dawn) Will Draw Fictional Women
03/12/2022, 3:33 PMjust use identical parameters to a previous session (if its still saved to drive) - a
a cat came into my house
03/12/2022, 6:59 PMFNF fleetway super sonic flatlined. - u
(".Lukus_bt.")
03/12/2022, 7:00 PMHmmm - u
(Dawn) Will Draw Fictional Women
03/12/2022, 7:04 PMshort wav! - f
F.B
03/12/2022, 10:18 PMCrying right now - u
(Dawn) Will Draw Fictional Women
03/12/2022, 10:19 PM>epoch 5 - u
(Dawn) Will Draw Fictional Women
03/12/2022, 10:19 PMdont even bother calling it now - u
(Dawn) Will Draw Fictional Women
03/12/2022, 10:19 PMwait til like 20 - f
F.B
03/12/2022, 10:19 PMIt's epoch 4 - u
(Dawn) Will Draw Fictional Women
03/12/2022, 10:19 PMpoint still stands - f
F.B
03/12/2022, 10:30 PMPain - f
F.B
03/12/2022, 10:31 PMInma wait - f
F.B
03/12/2022, 10:31 PMI don't wanna say that this will be non functional - f
F.B
03/12/2022, 11:04 PMIt failed - f
F.B
03/12/2022, 11:04 PMAs suspected - u
(Dawn) Will Draw Fictional Women
03/12/2022, 11:04 PMbad dataset - f
F.B
03/12/2022, 11:04 PM75 wavs - u
(Dawn) Will Draw Fictional Women
03/12/2022, 11:05 PMfind out which dont wanna align!! - a
Alphabet
03/12/2022, 11:53 PMFINALLY, got a complete list of subtitles for FF7 Remake and automated a way to grab and separate out all the audio files. - a
Alphabet
03/12/2022, 11:53 PMdoes anyone want a zip of the audio and subtitles transcription for anyone? I'm only going to do cloud, tifa, and aerith. So the rest are up for grabs - a
Alphabet
03/12/2022, 11:55 PMfor the main characters, there's several hours of audio and 1000+ wavs each. But the side characters have a few hours and hundreds of wavs too - m
mega b
03/12/2022, 11:56 PMwhoa nice! - a
Alphabet
03/13/2022, 6:53 AMtest speech with Aerith. For index0.wav, voice still sounds sratchy, but it's not bad. If I use my male voice as reference though, the outcome is much shittier. For the 2nd wav (index.wav), i use aerith's voice as a reference and it basically sounds perfect.