@gabrielck11 https://github.com/Hiroshiba/realtime-yukarin As dawn said, it'll probably be easiest to wait for nyvox, but if you want to check out a real time voice changer in the meantime you can check this if you have an nvidia gpu

a very good nvidia gpu at that

6gb vram

Oh damn i didn't notice that

The gpu i want to buy has 6gb vram I just checked it and it went back up in price though by like 100 dollars

I wonder if i can get it to work in colab...

I was about to train a new model and this happened

What happened?

What’s the fix?

Oh wait.

I think that’s the fix

Let me try

Oh wait, emoji 2.0.0 is out

You also have to update it

Update it how?

/content/project /usr/local/lib/python3.7/dist-packages/resampy/interpn.py:114: NumbaWarning: The TBB threading layer requires TBB version 2019.5 or later i.e., TBB_INTERFACE_VERSION >= 11005. Found TBB_INTERFACE_VERSION = 9107. The TBB threading layer is disabled. _resample_loop_p(x, t_out, interp_win, interp_delta, num_table, scale, y) Traceback (most recent call last): File "/usr/lib/python3.7/runpy.py", line 193, in _run_module_as_main "__main__", mod_spec) File "/usr/lib/python3.7/runpy.py", line 85, in _run_code exec(code, run_globals) File "/usr/local/lib/python3.7/dist-packages/uberduck_ml_dev/exec/train_tacotron2.py", line 6, in from ..trainer.tacotron2 import Tacotron2Trainer File "/usr/local/lib/python3.7/dist-packages/uberduck_ml_dev/trainer/tacotron2.py", line 63, in from ..models.torchmoji import TorchMojiInterface File "/usr/local/lib/python3.7/dist-packages/uberduck_ml_dev/models/torchmoji.py", line 100, in ALLOWED_EMOJIS = emoji.get_emoji_unicode_dict("en").values() File "/usr/local/lib/python3.7/dist-packages/emoji/__init__.py", line 97, in __getattr__ raise AttributeError("module 'emoji' has no attribute '%s'" % (varname, )) AttributeError: module 'emoji' has no attribute 'get_emoji_unicode_dict'

Someone help out.

-oh wait so uberduck updated and now i gotta remove the emoji clause

they merged my pr :)

!pip install --upgrade emoji

fixed it on my end

Guys?

I have some audios I would like to turn into a model for Uberduck

hecko ✗ [founder of h] said that I should ignore the training stuff

I don't know what I should do

ignore the training stuff in the guide because it's outdated

but either way you should first follow the dataset creation stuff

So the first video

I follow it PERFECTLY, right?

video??