Using the Model with Raspberry Pi 4B and a USB Mic Input

I have been trying to use this model with Raspberry Pi 4B while the input is from the USB Mic.

While writing inference for the model for Raspberry Pi, I am unable to take the input audio(preferably in the sliding widow format) and detect the command. I am only concerned with detecting the input here.

I did try multiple code snippets but ran into some issue. And the firmware code is a bit less commented for someone like me (beginner ) to get.

The issue names no files or tests; start by locating the Raspberry Pi inference entry point and the firmware code that handles audio input. Check how USB microphone data and sliding-window inference are currently represented, then verify that a Raspberry Pi can receive the audio and detect the command.

Written by the indexing model from the issue text.