what colab and what error

Traceback (most recent call last): File "main.py", line 54, in run_boostmonodepth(sample['ref_img_fi'], config['src_folder'], config['depth_folder']) File "/content/3d-photo-inpainting/boostmonodepth_utils.py", line 41, in run_boostmonodepth depth = imageio.imread(os.path.join(BOOST_BASE, BOOST_OUTPUTS, tgt_name)) File "/usr/local/lib/python3.7/dist-packages/imageio/core/functions.py", line 265, in imread reader = read(uri, format, "i", **kwargs) File "/usr/local/lib/python3.7/dist-packages/imageio/core/functions.py", line 172, in get_reader request = Request(uri, "r" + mode, **kwargs) File "/usr/local/lib/python3.7/dist-packages/imageio/core/request.py", line 124, in __init__ self._parse_uri(uri) File "/usr/local/lib/python3.7/dist-packages/imageio/core/request.py", line 260, in _parse_uri raise FileNotFoundError("No such file: '%s'" % fn) FileNotFoundError: No such file: '/content/3d-photo-inpainting/BoostingMonocularDepth/outputs/download_(4).jfif.png' 0% 0/2 [00:21<?, ?it/s]

oh haha

i have a better colab for that @User

You always have a better colab for everything don't you?

just run setup and run

is it only image url?

woooo

this is gonna take a while huh

You're getting there.

yeah i just didnt realize it was updating since i didnt scroll

thats reassuring

i cannot wait to torture my online friends with the cicero voice

The Colab downloaded the pretrained model. Is this normal or is something wrong?

??

thats. normal

Ok

Does Tacotron take flac files or do they all have to be converted to wav

This is neat than the original one...simplified!

Mailman's training is godlike

why does the discord lag keep happening

OOOOOOOooooOhhhhhhhhh fffffuuuuuuuuuukkkk nooohhh moooree moonnnee fucking nope noooo nah

it corrupted the mr narrator model and the colab fuck its self up

i think it restarted, maybe re-run cell 2? though maybe it's not prepared for being run when tacotron2 exists but the variables don't

moderate temptation to do lockpickinglawyer, i mean look at that treasure trove of data but it's his regular voice and my morals won't let me clone people's regular voices without permission and i'm not asking a 3m-subscriber-haver-and-also-lawyer for permission to enable theft of their identity plus tbh the meme potential is not even that high

unrelated: does anyone know whether there's any harm in training talknet models on super-long samples (let's say a minute even) from what i know talknet is convolutional instead of recurrent so it shouldn't care about sample length i don't think even if the duration predictor is recurrent by any chance, we're using reference audio anyway so it's just spectrograms and hifi-gan that matter and there might be the benefit of it learning to keep speech style consistent -except wait no the convolutionalness might not allow that

Hey Ya lyrics but the rest is AI generated

pretty good!