search

Kyubyong / dc_tts

A TensorFlow Implementation of DC-TTS: yet another text-to-speech model
Apache License 2.0
1.16k stars 370 forks source link

beginner question - error when running #71

Open reivilor opened 4 years ago

reivilor commented 4 years ago

Hello I m new to the field but i wanted to try voice synthesis during confinement im stuck with a computer that got 8 core Intel(R) Core(TM) i5-8250U CPU @ 1.60GHz

2020-04-14 02:13:50.492818: I tensorflow/core/platform/cpu_feature_guard.cc:137] Your CPU supports instructions that this TensorFlow binary was not compiled to use: SSE4.1 SSE4.2 AVX AVX2 FMA 0%| | 0/3 [00:00<?, ?b/s]Killed

well im stuck with this for now i think i need a GPU but i though i had a small one reading the doc of the proc, I ve very simple graphic card as its as barebone computer...

Thanks for any help

fathimad11 commented 4 years ago

How you solved this I also have this problem

fathimad11 commented 4 years ago

Well I solved this problem by changing the batch size to 2

reivilor commented 4 years ago

Wait, I ll try latter, I'm very interred. Thks!

fathimad11 notifications@github.com schrieb am Mi., 15. Apr. 2020, 4:54 PM:

Well I solved this problem by changing the batch size to 2

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/Kyubyong/dc_tts/issues/71#issuecomment-614123651, or unsubscribe https://github.com/notifications/unsubscribe-auth/AHMY6YCKEG3GPBUAA53CTMTRMXKEDANCNFSM4MHKWKHQ .

reivilor commented 4 years ago

Well thank you very much it seems to work for me also, Personally I reduce the number of sample in the database maybe you had to do the same also, like I let maybe 100 sample to try to get a first results, Maybe you have an insight to share about this, like how long did it take with 2 batch,

THXXX

Le mer. 15 avr. 2020 à 17:54, fathimad11 notifications@github.com a écrit :

Well I solved this problem by changing the batch size to 2

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/Kyubyong/dc_tts/issues/71#issuecomment-614123651, or unsubscribe https://github.com/notifications/unsubscribe-auth/AHMY6YCKEG3GPBUAA53CTMTRMXKEDANCNFSM4MHKWKHQ .

fathimad11 commented 4 years ago

When i started with batch size 2 , it generated a white allignment graph (all i did in my pc ), so I understood that it it not learning properly then i switch to google colab and batch size change to 16 , now training is going on in colab.

Current problem is I can't synthesize the sound it showing some error. If you synthesize the sound please help me..

fathimad11 commented 4 years ago

Everything solved... I got somewhat recognisable sound (my dataset have 32 audio sampe)

reivilor commented 4 years ago

Hey thank you for your answer I ll try again. Yes till 4 days in a row I have this never endind screen...

fathimad11 notifications@github.com schrieb am Di., 21. Apr. 2020, 8:57 AM:

Everything solved... I got somewhat recognisable sound (my dataset have 32 audio sampe)

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/Kyubyong/dc_tts/issues/71#issuecomment-617018938, or unsubscribe https://github.com/notifications/unsubscribe-auth/AHMY6YGNA4NINVUP6AG7FKLRNVGYFANCNFSM4MHKWKHQ .

reivilor commented 4 years ago

Okay thank you I ll use colab too I did not know this tool

Olivier R olivier.radziminski@gmail.com schrieb am Di., 21. Apr. 2020, 1:43 PM:

Hey thank you for your answer I ll try again. Yes till 4 days in a row I have this never endind screen...

fathimad11 notifications@github.com schrieb am Di., 21. Apr. 2020, 8:57 AM:

Everything solved... I got somewhat recognisable sound (my dataset have 32 audio sampe)

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/Kyubyong/dc_tts/issues/71#issuecomment-617018938, or unsubscribe https://github.com/notifications/unsubscribe-auth/AHMY6YGNA4NINVUP6AG7FKLRNVGYFANCNFSM4MHKWKHQ .

Crazyjoedevola commented 4 years ago

n i started with batch size 2 , it generated a white allignment graph (all i did in my pc ), so I understood that it it not learning properly then i switch to google colab and batch size change to 16 , now training is going on in colab.

Current problem is I can't synthesize the sound it showing some error. If you synthesize the sound please help

Hi, can you please let me know how to get this working on colab?