Скачать или смотреть Kyutai TTS LOCAL Test & Install (A VERY Expressive Voice Model)

Timestamps:
00:00 - Intro
01:03 - First Look
01:55 - Technical Look
03:25 - Pre Install Notes
04:25 - Local Install Guide
06:49 - First Test
07:35 - Gradio Test Script
08:26 - Dialogue Testing
11:15 - Style Testing
13:32 - Eccentric Testing
15:38 - Whisper Testing (Sensory Warning)
16:43 - Watercooler Scene Testing
20:04 - Closing Thoughts

AI Integration Consulting: https://bijanbowen.com
Discord for AI Discussions:   / discord  

In this video, we take a look at the newly released Kyutai TTS model — a multilingual, multi-style speech synthesis system developed by Kyutai Labs. The model supports a wide range of expressive voice styles and is positioned to compete with the best in both open-source and closed-source TTS for quality, latency, and robustness.

We start with a quick technical overview, followed by a full local install guide, including pre-install tips and dependencies. Once installed, we test the model using both CLI and a custom Gradio interface, which you can launch in-browser for easy experimentation. The demo includes dialogue testing, stylistic variation, whispers, and even some absurd, expressive tests to explore how far the voice control features can go.

Model GitHub: https://github.com/kyutai-labs/delaye...
Model on Hugging Face: https://huggingface.co/kyutai/tts-1.6...
Gradio Web UI Script: https://gist.github.com/OminousIndust...