I thought I had to do something with the the transcript

yeah just use the same one i think lol

like divide it into two different ones

OK

> RuntimeError: CUDA out of memory. Tried to allocate 86.00 MiB (GPU 0; 15.78 GiB total capacity; 13.62 GiB already allocated; 64.75 MiB free; 14.42 GiB reserved in total by PyTorch)

GOOGLE COLAB.

YOU HAVE A HUGE GPU

AND A LOT OF RAM

64 gigs of ram

jesus

I think 32 batches is too big.

I'll lower it to 28.

I hate Tacotron sometimes.

HONESTLY MOOD, I've been trying to get it to read my files for a solid 24 hours now dfhdffghfg

XD

Tacotron hates me

protip! usage limits can be avoided with spare google accounts... which I apparently need to use because I hit the limit trying to troubleshoot the errors that keep popping up

Training Strong Bad.

wooooooooo

All we need left is a Tycho Brahe model and then we have the Poker Night quartet

Hell yeah

I think Strong Bad's ready to be submitted.

Still need to do a bit more training but here's a sneak peak for Omni-Man

now make his do cave johnson lines

Oh shit I didnt even think to add some in there

I did add in quotes from whiplash though

Just gonna add those right in

heeee's finally training aaaaaaaa it's only working w the first batch of 20 clips rn but this is the furthest I've gotten yet! q u q

AND ALSO to answer this for those of you who are new, the "clipper train" bit is BASICALLY CoLab saying "yo I don't know where you put your .txt but I can't find it" I'm thinking it may be a glitch bc I changed the parameters a few times and reverted back to my first input and apparently it fixed it but it may also be sheer luck saying "aw lookit this poor fool still trying to synthesize speech 24 hours later, they may have little a mercy,, as a treate"