ar

Contextual grapheme-to-phoneme for Arabic, trained with grafon on top of UBC-NLP/MARBERTv2.

from grafon import Phonemizer

phonemize = Phonemizer.from_pretrained("grafon-g2p/ar")
print(phonemize('مرحبا بالعالم، ما رأيك في النموذج؟'))

Phonemes: abdfhijklmnpqrstuvwxzðħɡɣʃʒʔʕθˈːˤ. Anything outside the letters ءآأؤإئابةتثجحخدذرزسشصضطظعغفقكلمنهوىيپچڤگـ passes through as written.