your directory is probably not in the tacotron2 repo

%cd /content/tacotron2/

I'm using the inference notebook.

adding a new model to the drop-down

hmm i wonder how much time is spent on validation during training like how much faster could it go if the validation filelist was limited to 100 clips or something especially for super-massive datasets -though actually for those we'd want to have separate val clips that aren't in training anyway

can I expect to get confirmation/denial/feedback on my model soon? I submitted it about a month ago now

I'd recommend a ticket, go to #813143605862203442 and at the bottom you'll see a ticket thing

thanks

help

i said help

change folder_name in line 5 to wavs

got the fine tune pepper to work

he hasn't seem to change eh

wdym line 5

nvm i got it working

but now the only last ploblem is this

can you scroll down abit more so i can see the full output

why does it say wavs instead of folder in your notebook

i said scroll down not up

ok

are you sure the folder path you gave is not empty

it is empty

i forgot to add one

@User quacked: cococcsecsa.

can i use numbers in my dataset txt?

if you spell them out (as in "one" instead of "1"), yes

Sid, it automatically replaces 1, 2, ... with one, two, ... inside the colab

iirc

i still think its better because of larger numbers