I assume your model file is broken

Oh dear...

then just repeat the process

And don't end it prematurely this time?

you can specify the number of epochs

just put it to 40 epochs or so

and it'll be fine

That's actually what I wonder, what the heck is epoch?

Ah, number of times the ai looks through the files, I see...

Google what an epoch is lol

Yeah

either way, wouldn't the voice come out more fuzzy if the epoch is low?

But you terminated after 34 epochs

So no

Oh heavens no, that pic I sent above isn't the model stopped, it's of the new one I'm trying to do.

It's still going as we speak

the default in the colab is set redicolously high

so you've got the reduce it

or stop it

But that's the thing.

If I do stop it prematurely, isn't that running the risk of certain attributes of it just... not appearing?

just limit the epochs before starting the training

or take a checkpoint as your model

Like the mysteriously missing central directory... what could be the reason for that if the checker within the trainer module said everything was okay?

I don't know enough about your context

ML is a somewhat difficult topic and doing without carefully following instructions is close to impossible

But that IS the context? The audio is set at the right rate and is mono, the module knows the list file and the audio files, and the checker for making sure nothing is screwed up came back with saying everything within the files were right.

You corrupted your model file

Or tooks something else as your model

I beg your pardon?