How to create CMUDict for Portuguese Language?

NVIDIA / flowtron

Flowtron is an auto-regressive flow-based generative network for text to speech synthesis with control over speech variation and style transfer

https://nv-adlr.github.io/Flowtron

Apache License 2.0

887 stars 177 forks source link

How to create CMUDict for Portuguese Language? #98

Closed custodiogabriel closed 3 years ago

custodiogabriel commented 3 years ago

I found this dictionary https://github.com/NVIDIA/flowtron/issues/75, but how do I convert to CMUDict? Heteronymus and CMUDict

rafaelvalle commented 3 years ago

Don't convert it to CMUDict. Use the set of phonemes in the Portuguese dictionary, which is probably using IPA. Replace the set of phoneme characters from ARPABET to the set of phoneme characters in the portuguese phoneme dictionary. https://github.com/NVIDIA/flowtron/blob/master/text/symbols.py#L17 Share samples when you have some!

custodiogabriel commented 3 years ago

It's ok! I'm trying to use the Mozilla PTBR data set in this code :) I sent an invitation on your Linkedin

rafaelvalle commented 3 years ago

closing due to inactivity.