mahwizzzz

/

orpheus-urdu-tts

like 5

Dear mahwizzzz and the Orpheus Urdu TTS Community,

My name is Zaryab Fayaz, and I am writing from Ferozewala, Pakistan. I am deeply impressed by your exceptional work on the Orpheus Urdu TTS model, which is a vital step forward for Urdu digital accessibility.

I have thoroughly reviewed the Model Card. While I acknowledge that this is an "experimental model not recommended for production use", I believe that optimizing it for mobile is the most critical next experimental step for maximizing its impact.

The Intended Use section explicitly mentions "speaking to people with hearing impairment," "text reading for education purposes," and "conversational Al".

1. A Personal Necessity for Accessible Communication

My request is rooted in personal necessity:

I am 20 years old and live with a

stuttering condition that makes natural,

clear speech exceptionally difficult.

I also face a physical disability that requires me to operate my mobile device using my nose.

For individuals like me, a high-quality Urdu TTS solution is not merely a convenience; it is a fundamental tool for effective self-expression and social participation, directly fulfilling the accessibility goal stated in your Model Card.

2. Addressing the Lack of Urdu Content Creation Resources

Beyond personal communication, there is a severe lack of accessible and efficient tools for Urdu content creators. Your model is a crucial resource, yet its current deployment. size (6.63 GB) and server-bound format restrict its use.

My Request for Dual Deployment Options (The Next Experiment)

To move this experiment into the hands of real users and test its utility under varied conditions, I humbly request that you consider one of two deployment options:

1. Offline Solution (Preferred): The creation and release of a mobile-optimized version in ONNX or TensorFlow Lite (TFLite) format, ideally with Quantization applied to allow for true offline usage on mobile devices. This would be a groundbreaking experiment in edge device TTS.

2. Online Solution (Alternative): The deployment of a hosted API (Application Programming Interface) for the model. This would allow developers to integrate the TTS capability into lightweight apps, ensuring reliable (though online) access to the service.

My Commitment to Overcoming Barriers Despite these physical and educational barriers, I have never let them hinder my drive for excellence. I have successfully completed numerous challenging, non-credit courses through Coursera's Financial Aid program from world-class universities, demonstrating my deep commitment to self-development:

Financial Markets (Yale University, with Honors)

Cryptography I (Stanford University) Writing in the Sciences (Stanford University, with Honors) Developing a Systems Mindset (University of Colorado Boulder)

Your contribution, through either an optimized file or an accessible API, will directly support the accessibility goals of this remarkable model and empower individuals facing communication challenges worldwide.

Thank you for your generosity, vision, and consideration. I eagerly await any news regarding a more accessible release.

Sincerely,

Zaryab Fayaz
Ferozewala, Punjab, Pakistan

Connect with me:

Linkedin Profile:

https://www.linkedin.com/in/zaryab-fayaz-670650297?utm_source=share&utm_campaign=share_via&utm_content=profile&utm_medium=android_app