Datasets:
need info about the dataset
Hi Mahta!
Thanks for sharing this dataset.
Is the dataset multi-speaker or mono? how many hours of voice does it have?
Has a model been developed on this dataset yet? I would appreciate it if you could introduce it.
Hi Fatemeh!
Thank you for your interest in the dataset! I really appreciate your questions—they helped me notice that some key details were missing from the README. I’ve just updated it to include:
- Confirmation that the dataset is single-speaker with 102 hours of audio.
- A link to the Tacotron2 model trained on this dataset.
You can also take a look at our NAACL 2025 paper "ManaTTS Persian: A Recipe for Creating TTS Datasets" for details about the dataset creation and model evaluations. Let me know if you have any other questions or need further details—I’m happy to help!
Thanks for your answer.
I want to trian a tts model on this dataset. I would be happy to share the results with you.
That’s great to hear! We’re excited to see your results—thank you so much for sharing them with the community.