Description
Indic Parler-TTS is a multilingual text-to-speech model designed to support 21 languages, including English and a variety of Indic languages such as Hindi, Bengali, Tamil, and more. Developed through a collaboration between Hugging Face and AI4Bharat, this model is a fine-tuned version of Indic Parler-TTS Pretrained, trained on a dataset of 1,806 hours.
Read More
The model provides extensive language support, officially covering languages like Assamese, Bengali, Bodo, Dogri, Kannada, Malayalam, Marathi, Sanskrit, Nepali, and Tamil. It also offers unofficial support for languages like Chhattisgarhi, Kashmiri, and Punjabi. Indic Parler-TTS features 69 unique voices across supported languages, optimized for naturalness and intelligibility.
Emotion rendering is available for 10 languages, allowing users to specify emotion-specific prompts such as Command, Anger, Narration, and more. The model supports Indian English accents and allows customization for other accents using style transfer techniques.
Users can control various speech characteristics using descriptive captions, including background noise, reverberation, expressivity, pitch, speaking rate, and voice quality. The model adapts automatically to the language detected in the prompt, ensuring seamless language switching.
Indic Parler-TTS has been evaluated using a MOS-like framework, demonstrating exceptional performance in generating natural and intelligible speech, particularly for native speakers of Indian languages. The model is permissively licensed under the Apache 2.0 license, making it accessible for various applications in regional language technologies.
- CategoryText to Speech
- Pricing ModelFree
- DeploymentCloud
- PlatformsWeb
Indic Parler-TTS Highlights
Multilingual support for 21 languages
69 unique voices across languages
Emotion-specific prompts for 10 languages
Indian English accent support
Customizable speech characteristics
Automatic language detection
MOS-like evaluation framework
Apache 2.0 license
Getting Started with Indic Parler-TTS
Access page: Visit the Hugging Face repository
Load model: Download the Indic Parler-TTS model
Configure environment: Set up necessary dependencies
Integrate: Use the model in your application
Fine-tune: Customize the model for specific needs
Indic Parler-TTS's Use Cases
- Multilingual TTS
- Emotion Rendering
- Accent Customization
- Speaker Diversity
- Regional Language Tech
