How to run the models natively in a python file

[MichaelG]

Hello,

I have a bunch of audio files stored locally which I’d like to run the model on and compare the output from the model to the known value (basically to check the accuracy of the model).

Right now I’m using os to run things but I’m not sure that’s the best way. I’d like to retrieve the value after the model has finished estimation and then compare the output to the true value but it seems I only get a None type:

import os

val = os.system("deepspeech --model deepspeech-0.9.3-models.pbmm --scorer deepspeech-0.9.3-models.scorer --audio hello_world.wav")

print("val + ", val)

Any ideas how to get the output from the model so I can compare to the actual text

[lissyx ((slow to reply) [NOT PROVIDING SUPPORT])]

@MichaelG

I have a bunch of audio files stored locally which I’d like to run the model on and compare the output from the model to the known value (basically to check the accuracy of the model).

You should just follow the training docs and use evaluate.py for that usecase.

BTW Python bindings is documented: Python API Usage example — DeepSpeech 0.9.3 documentation