aminbaig

Any update on this...? I have been having the same issue.

Is there a way to even bring the resolution lower then HD? For example if we were to reduce it to SD instead of HD or Full HD can we...

I think I found the issue. in my urdu.py I changed the text=parts[1] and now its preprocessing. Is this the right approach?

Thanks @begeekmyfriend. One question, my card is GTX 750 Ti which support compute capability 5.0 Will that be an issue?

@begeekmyfriend I have upgraded my cua to 10, cuDNN to 7.5 my tensorflow-gpu is 1.13 python is 3.5 but still getting error when trying to train. I have attached the...

@begeekmyfriend I tried that as well...even reduced the max_mel_frames to 10. Following is the error text attached. [error2.txt](https://github.com/keithito/tacotron/files/2976203/error2.txt) I have also attached my hyperparameters file. [hparams.py.zip](https://github.com/keithito/tacotron/files/2976208/hparams.py.zip)

I did as you advised...still getting the same error

my scipy version is 0.19.0

Hi Tayyaba, First change your cleaners to "basic_cleaners", secondly, have you updated the symbols with unique characters from the urdu text? Last but most important...your training data set is too...

Thanks Sahar, I have picked up the project again. Yes it and thank you for your response. I am now having issues training the data as I get OOM error...