also should i leave the learning rate decay thing on or off

cuz its tricky to understand this part

I say leave it on. Ive no real way of accurate dataset measure but I would say over 5+ maybe leave as-is??

I've already given him instructions in DMs

Ok

fawk

was working fine hour ago

this is the synthesis notebook

NameError Traceback (most recent call last) in 38 try: ---> 39 initialized 40 except NameError: NameError: name 'initialized' is not defined During handling of the above exception, another exception occurred: TypeError Traceback (most recent call last) 3 frames /content/hifi-gan/stft.py in __init__(self, filter_length, hop_length, win_length, window) 65 # get window and zero center pad it to filter_length 66 fft_window = get_window(window, win_length, fftbins=True) ---> 67 fft_window = pad_center(fft_window, filter_length) 68 fft_window = torch.from_numpy(fft_window).float() 69 TypeError: pad_center() takes 1 positional argument but 2 were given

^ Having this error too.

the synthesis notebooks seem to be breaking more frequently lately

weird

Also affected the training one

At least you can use Fallback Runtime

Fallback Runtime is a divine gift from God

use the updated nb

using the links posted in #841437191073955920

thanks for fixing it bro i dunno what we would do without you!

fr tho

^

hi lol colab is throwing me an error saying this file doesnt exist but it does

theres no spaces

is colab on drugs

it clearly exists in the transcription and in the voice clips

its pipeline tacotron

@hecko u made the note book right? may you help me like do the magic on the notebook so it can do the correct thing?

Did you check your dataset?

Or something.

I remember getting that error.

yh i did