Does it support other languages?

#1
by sadafwalliyani - opened

Does it support other languages?

Collabora org

We have English and Polish. Support for French and German is in the works, should be release in the next few weeks.

Which languages are you interested in?

Is there going to be any support for Hungarian? (I know datasets are scarce for Hungarian, but there is at least one, I think a full reading of a book performed by one voice only. The dataset is called "hungarian-single-speaker-tts".)

Collabora org

I think Hungarian may be challenging if we don’t find good data. From what I know it is very different from other European languages so it’s unclear how much knowledge transfer we can expect from the model.

But we can give it a try. We are working on specifying the dataset format that could be easily ingested into the training process so more people could help gather the datasets.

Waiting for the French language to be added to this demo soon 🤗

can you add Arabic language, please

Great work on this! Can you please consider adding Indonesian language?

We are working on specifying the dataset format that could be easily ingested into the training process so more people could help gather the datasets.

This is mostly anticipated feature. Looking forward to contributing with some languages.

Collabora org

Hey,

I updated the Collab notebook (linked from the README) that showcases a preliminary model release with 8 languages.

One thing that would help a lot is if someone downloaded and converted more multilingual data to WebDataset.

Sign up or log in to comment