toebee

3 ms·

61.
by toebee 1y ago
We will work on a quantized version of the model, so hopefully you will be able to run it soon! We've seen Bark from Suno go from 16GB requirement -> 4GB requirement + running on CPUs. Won't be too hard, just need some time to
62.
by toebee 1y ago
Sounds awesome! I think it won't be very hard to run it using output streaming, although that might require beefier GPUs. Give us an email and we can talk more - nari.ai.contact at gmail dot com. It's way past bedtime where I live
63.
by toebee 1y ago
We're envisioning a platform with a social aspect, so that is the biggest difference. Also, bigger models! We are aware of the fact that you do not need to create a venv when using pre-existing uv. Just added it for people spinning up
64.
by toebee 1y ago
Interesting. I haven't thought of that problem before. I'm guessing a large enough audio dataset for medical terminology does not exist publicly. But AFAIK, even if you have just a few hours of audio containing specific terminolog
65.
by toebee 1y ago
Thanks for the heads-up! We weren’t aware of the GNOME Dia project. Since we focus on speech AI, we’ll make sure to clarify that distinction.
66.
Show HN: Dia, an open-weights TTS model for generating realistic dialogue(github.com)
652 points by toebee 1y ago|190 comments
67.
by toebee 1y ago
Hey HN! We’re Toby and Jay, creators of Dia. Dia is 1.6B parameter open-weights model that generates dialogue directly from a transcript. Unlike TTS models that generate each speaker turn and stitch them together, Dia generates the entire c