Regarding Inter frame processing in training time v/s Real time inferencing. #7
Description
Hi,
I am trying to understand about inter frame processing. my understanding is that for every frequency inter frame LSTM go through all the time and returns h0, c0 after this for 2nd frequency the same process repeats and this process continuous for all the frequency bins.
In Real time inferencing we have only one frame, now 1st frequency is passed to the Lstm, since there is no next frame, the Lstm looks into 2nd frequency. how it is equal to the training. we might get different outputs as we are showing differently.
The above is my understanding of inter frame LSTM.
please guide me, if I am wrong, I wanted to know how we are going to implement inter frame processing without any delay in realtime.