[HN Gopher] Llama 3.1 Omni Model
___________________________________________________________________
Llama 3.1 Omni Model
Author : taikon
Score : 106 points
Date : 2024-09-18 16:42 UTC (6 hours ago)
(HTM) web link (github.com)
(TXT) w3m dump (github.com)
| nickthegreek wrote:
| The speed looks very nice. I just recently setup LMStudio +
| AnythingLLM to try out local voice chat and its still a little
| slower than I'd like but the PiperTTS voices are nicer than this.
| opdahl wrote:
| Any demos showcasing it's performance?
| potatoman22 wrote:
| There's one on Huggingface
| https://huggingface.co/ICTNLP/Llama-3.1-8B-Omni
| opdahl wrote:
| Thank you.
|
| Obviously it doesn't sound human but that's extremely
| impressive for an 8B model. Compared to the Moshi model also
| on the front page now, this model seems to be more coherent,
| but maybe less conversational?
| twobitshifter wrote:
| There is a demo video on the page
| dingdingdang wrote:
| Does any of the model-runners support this? Ollama, LM Studio,
| llama.cpp?
| LorenDB wrote:
| The TTS voice in the demo clip sounds remarkably like Ellen
| McLain (Valve voice actor).
|
| https://en.m.wikipedia.org/wiki/Ellen_McLain
| londons_explore wrote:
| Can this play sounds that can't be represented in text? Ie. "make
| the noise a chicken makes"
| hansenliang wrote:
| asking the real questions
| indigodaddy wrote:
| Very interesting question actually
___________________________________________________________________
(page generated 2024-09-18 23:00 UTC)