lammoh

Hi, I noticed that you released a smaller pretrained model (~190M), it's different from the one mentioned in the paper (1B), I'd like to know when do you expect to...

I'm fine-tuning the released checkpoint (~190M) on a new language for plain_tts task. The initial results so far look good as the voice sounds natural, but still not satisfying. I...

Warning: Grad strides do not match bucket view strides.

1

I'm training MeloTTS on my custom dataset + new language ``` /opt/conda/envs/melotts/lib/python3.9/site-packages/torch/autograd/graph.py:768: UserWarning: Grad strides do not match bucket view strides. This may indicate grad was not created according to...

gradio 4.41 doesn't play audio in IOS

3

### Describe the bug gradio 4.41 doesn't play audio in IOS downgrading to 4.40 solve the issue ### Have you searched existing issues? 🔎 - [X] I have searched and...

Two phones in Arabic are mapped to the same symbol

2

I noticed that "ظ" (ðˤ) and "ذ" (ð) are both mapped to "ð" ``` from phonemizer.backend import BACKENDS g2p = BACKENDS['espeak'](language='ar', words_mismatch='warn', preserve_punctuation=False) g2p.phonemize(["ظ"], separator=Separator(phone=None, word=' ', syllable='|'), strip=True) #...