this project don't works, bad porject.

[Chancing0]

while(1): means that it would never stop ,i have been train for few days ,but just never get finished. And the code was also wrong in some ways. The train data would be always the initial 64 audios.and train it with while(true), so we'd always training our model with the same initial 64 audios In face , X,Y=nextbatch should be get inside of the whlie loop,Then each epoch can get the next 64 audio'data to training modle.

[rain-1]

I had the same problem

I changed it to this:

for i in range(0, training_iters):

[rain-1]

https://github.com/llSourcell/tensorflow_speech_recognition_demo/pull/14

[rain-1]

https://github.com/llSourcell/tensorflow_speech_recognition_demo/issues/12

[hcchengithub]

@Chancing0 Thank you so much! OMG you are right!!!

the train data would be always the initial 64 audio.and train it with while(true), what i say , X,Y=nextbatch should be inside of the whlie loop,then each epoch can recive the next 64 audio'data to train

Have you fixed all the problems? May I have your GitHub address if you have?

[Chancing0]

@hcchengithub you could see my little porject,,https://github.com/Chancing0/LSTM-Speech-Reconization-Project
Hope i could help you.

[hcchengithub]

This is the original on pannous' repo: https://github.com/pannous/tensorflow-speech-recognition/blob/master/lstm-tflearn.py