Cloned Voice for AI Agents using Sesame CSM 1B - A Quick Demo

This conversation was created using the Apache 2.0 licensed CSM 1B by Sesame Labs. It is a Conversational Speech Model.

It has been trained on 1 MILLION hours of data, is contextually aware, is emotionally intelligent speech. The model also support Voice cloning as seen in this demo. It is based on llama architecture & Mimi like decoder.

https://github.com/SesameAILabs/csm
https://huggingface.co/sesame/csm-1b