#Training error
1 messages · Page 1 of 1 (latest)
basically... pretrains screwed up
it couldn't find one of pretrain files
just use original atp
Oh geez, I was about to try TITAN model. But nevermind I guess
have u run preprocess and f0 without issues?
even with TITAN should run fine
Yeah, everything went great, except for that Training part...
I still don't understand what it meant by "No such file or directory" part on the last line
I always used only alphanumeric on the model name, and no issues at all
I'm sorry, but could you explain that to me? I don't think I understand
im kinda bad at explaining anything that tech savvy people should know, sorry
It's fine. I'll learn more if I keep doing this thing every day. Maybe I should try using Applio because Google Colab can't be reliable sometimes
basically... alphabet and numbers.
only use alphabet & numbers, nothing more
if you have decent Nvidia gpu to train locally
Ahh I get it now. I always use alphabet for naming my models. So, I shouldn't have encountered any issues. Perhaps it's just internal issues or something. Nothing stays stable I guess
I'm using NVIDIA GeForce RTX 3050 Laptop GPU. Idk if that'd do
4 gb vram right? seems not really feasible, where batch size 4 as bare minimum needs 6 gb vram to run without severe performance bottleneck
aww man that's so lame. I'll try Colab again tomorrow
you can switch to another account if u want
Thanks for helping me! I'll keep trying to resolve my issues
That means it tried to read that file (filelist.txt) but couldn't find it, I think the pretrains are bugged as @misty hazel pointed out. I couldn't use the Ov2 32k or TITAN earlier today
Actually might be RVC disconnected issue, because later I downloaded the titan pretrains and was able to use it locally with applio