The trained model checkpoints are intended for use in TTS applications, research, and development, providing a resource for those looking to incorporate high-quality Ukrainian speech synthesis into their projects.
To utilize these checkpoints in your projects, refer to the GitHub repository for implementation details and examples:
Training Dataset: Synthetic dataset comprising 1,388 audio files and their corresponding texts, totaling 2 hours and 20 minutes of speech.
Objective: The model aims to demonstrate the efficacy of synthetic datasets in training robust and versatile TTS systems.
License
The checkpoints are distributed under the MIT License, allowing for both academic and commercial use, provided that proper credit is given.
Citation
If you find these checkpoints useful in your research or application, please consider citing the repository as follows:
@misc{pflowtts_uk_elevenlabs_checkpoints,
author = {@skypro1111},
title = {pflowtts_uk_elevenlabs Checkpoints},
year = {2024},
publisher = {Hugging Face},
journal = {Hugging Face Repository},
howpublished = {\url{https://huggingface.co/skypro1111/pyflowtts_uk_elevenlabs}}
}
Acknowledgments
Special thanks to the services provided by ChatGPT-4 and ElevenLabs.io for making the generation of the synthetic dataset possible, thereby facilitating the development of this TTS model.
