---
title: "IISc's Vāgdhenu Opens Sanskrit Chant TTS With 4.6 Expert Score"
subtitle: "A single-speaker Sanskrit chant text-to-speech system fine-tuned from IndicF5 hits MOS ~4.6 and cracks retroflex conjuncts that stumped earlier architectures."
url: "https://alphasignal.ai/news/iisc-s-vagdhenu-opens-sanskrit-chant-tts-with-4-6-expert-score"
author: "AlphaSignal Newsroom"
author_url: "https://alphasignal.ai/editorial-team"
company: "AlphaSignal"
company_url: "https://alphasignal.ai/company/alphasignal"
published: "2026-06-28T15:29:15.000Z"
updated: "2026-09-18T02:40:49.715Z"
type: "Model"
topics:
  - "Audio"
  - "Open Source"
subtopics:
  - "Text To Speech"
  - "Voice Cloning"
upvotes: 651
source_url: "https://huggingface.co/prathoshap/vagdhenu"
comments: 0
truncated: true
---

# IISc's Vāgdhenu Opens Sanskrit Chant TTS With 4.6 Expert Score

> A single-speaker Sanskrit chant text-to-speech system fine-tuned from IndicF5 hits MOS ~4.6 and cracks retroflex conjuncts that stumped earlier architectures.

# Vāgdhenu brings meter-aware Sanskrit chant to open TTS

[Vāgdhenu](https://huggingface.co/prathoshap/vagdhenu), an Apache 2.0 text-to-speech system for Sanskrit chant, is now available on Hugging Face with its code, weights, and training data public. It recites classical ślokas in a traditional pārāyaṇa style, preserving metrical duration and melodic contour. Prof. Prathosh at the Indian Institute of Science, Bengaluru, built the system on [AI4Bharat’s IndicF5](https://huggingface.co/ai4bharat/IndicF5) and F5-TTS, paired with a fine-tuned NVIDIA BigVGAN-v2 vocoder.

The project reports a mean opinion score of about 4.6 from one expert listener and 100% correct rendering of the tested conjunct clusters, including difficult retroflex combinations such as ṣṭ and ḍḍh. Pronunciation errors in these clusters can change the recited form of a verse. With one listener and no stated cluster-test-set size, the figures provide project-level evidence and leave broader benchmarking open.

## A flow model, steered by examples

The acoustic backbone is a roughly 337-million-parameter diffusion transformer trained with optimal-transport conditional flow matching. It learns a velocity field that moves noise toward a target mel spectrogram, a time-frequency representation of audio, while conditioning on text and a reference recording. The optimal-transport formulation encourages relatively direct generation paths, which can reduce sampling work and stabilize training. BigVGAN then converts the generated spectrogram into a waveform.

_Preview only. Full text requires an AlphaSignal Pro account._

## Sources

- [prathoshap/vagdhenu on Hugging Face](https://huggingface.co/prathoshap/vagdhenu)
- [Vāgdhenu GitHub repository](https://github.com/prathoshap/vagdhenu)
- [prathoshap/vagdhenu-data dataset](https://huggingface.co/datasets/prathoshap/vagdhenu-data)
- [AI4Bharat IndicF5 base model](https://huggingface.co/ai4bharat/IndicF5)