Audio resources

• There are following open-source datasets currently available on Odia speech.

Dataset	License	Need to be cleaned?	Estimated hours of audio duration	Note
Mozilla Common Voice	CC-0	No	11hrs (out of which ~2hrs verified)	Odia text to speech/Speech to text corpus. Our team actively contribute on this project.
Odia Pronunciations	CC-BY-SA-4.0	No	20,000+ words/phrases

---

To cite this resource list, please use:

@misc{OdiaNLP,
    author       = {Soumendra Kumar Sahoo},
    title        = {Audio resources by Odia NLP},
    howpublished = {\url{https://www.mte2o.com/}},
    year         = {2021}
}

---

Last update: 2023-03-27