←back to thread

Show HN: Dia, an open-weights TTS model for generating realistic dialogue

(github.com)
652 points toebee | 4 comments | 21 Apr 25 17:07 UTC | HN request time: 6.269s | source
Show context
ivape ◴[21 Apr 25 17:34 UTC] No.43754400[source]▶
Darn, don't have the appropriate hardware.

The full version of Dia requires around 10GB of VRAM to run.

If you have a 16gb of VRAM, I guess you could pair this with a 3B param model along side it, or really probably only 1B param with reasonable context window.

replies(1): >>43754481 #
1. toebee ◴[21 Apr 25 17:43 UTC] No.43754481[source]▶
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 work on it.

replies(1): >>43754581 #
2. ivape ◴[21 Apr 25 17:54 UTC] No.43754581[source]▶
No doubt, these TTS models locally are what I'm looking for because I'm so done typing and reading :)
replies(1): >>43758083 #
3. toebee ◴[22 Apr 25 00:50 UTC] No.43758083[source]▶
You can try it now on https://huggingface.co/spaces/nari-labs/Dia-1.6B !!!
replies(1): >>43766019 #
4. ivape ◴[22 Apr 25 20:36 UTC] No.43766019{3}[source]▶
Woah, you guys are on it!