hi
I used “simple audio” sample.
every things fine, data are 1 seconds for train and input data for test result should be 1 second too.
if i send 1 recorded audio in 2 seconds, system cannot find result
how can i do that ?
how about longer audio files? how about Sentence?
Hai @rahmani2022 & Welcome to the Tensorflow forum.
You’re refering to Simple audio recognition: Recognizing keywords tutorial put together by the Tensorflow, right? Looking quickly into it, one word among of fixed / pre-defined list of 10 words (phonemes) is presented to the model which outputs the probabilities that word is each of the 10 possible words. So it is not really “scalable” to longer inputs.
Maybe Speech Recognition: a review of the different deep learning approaches will provide you guidance on the next step in your endeavor?