Titu TTS (Text-to-Speech) is an open-source initiative by Hishab, designed to enhance text-to-speech synthesis, particularly for the Bangla language. This project is a key component of the broader TituLM effort, which focuses on pushing the frontiers of text-to-speech technology for product innovation.
At Hishab, we are committed to refining text-to-speech technologies through a robust pipeline of pretraining and fine-tuning models on various TTS tasks. While Titu TTS primarily targets the Bangla language, its adaptable technology framework allows for potential extensions to other languages in the future.
Our objective is to develop precise and reliable text-to-speech models that can be seamlessly integrated into diverse applications, thereby improving accessibility and enriching the user experience across various sectors.
The deployment is designed to support both streaming and inference functionalities and is integrated with RESTAPI as well as Web Socket for versatile deployment options.
Paper:
@misc{Titu TTS,
title={Titu TTS for Bangla Text to Speech Generation Model},
author={Hishab Technologies LTD},
howpublished={},
year={2024}
}