How to use fluxions/vui with Vui:
# !pip install git+https://github.com/fluxions-ai/vui import torchaudio from vui.inference import render from vui.model import Vui, model = Vui.from_pretrained().cuda() waveform = render( model, "Hey, here is some random stuff, usually something quite long as the shorter the text the less likely the model can cope!", ) print(waveform.shape) torchaudio.save("out.opus", waveform[0], 22050)
17Ge