search

yoosif0 / arabic-tacotron-tts

End to end Arabic TTS system based on tacotron
MIT License
116 stars 35 forks source link

No alignment for Urdu #8

Open AownMohammad opened 4 years ago

AownMohammad commented 4 years ago

Hi, I have used this model to train on my speech Urdu dataset. It contains 10000 .wav files of 15 hours speech dataset. Average file is size 5.4 sec. I used default parameters and trained for 50000 steps. I used transliterated labels and transliteration_cleaner. Here is the alignment. step-50000-align

tayyabvohra commented 4 years ago

Can you please share the synthesize audio file .

AownMohammad commented 4 years ago

Thanks for replying. Here are the synthesized audio files at 50000 steps. www.aown.me/eval-50000-0.wav www.aown.me/eval-50000-1.wav www.aown.me/eval-50000-2.wav Here is training sample www.aown.me/45.wav

tayyabvohra commented 4 years ago

@AownMohammad the same problem has occured with me I have iterate over 500K steps and my loss is 0.06 but still the problem is same.

AownMohammad commented 4 years ago

@tayyabvohra I have been trying with different settings but none is working. If it the model can work on Arabic then it should work on Urdu too.

tayyabvohra commented 4 years ago

@AownMohammad I think we should tune the hyper parameters according to it.

AownMohammad commented 4 years ago

@tayyabvohra But how? training takes alot of time. Training again and again with different hyprams would become impossible.