Hello, I was able to compile an edge impulse model and used the audiomoth-edgeimpulse repo to flash onto my audiomoth. it does appear to be working, I'm just confused about a few settings and how to interpret the results.

My main question is- Edge Impulse requires at least 2 classes to train an acoustic classifier. I have one class being my bird of interest, and my second class is "background noise" of either other birds in the area or just regular background noise. But I was unable to see where in the audiomoth firmware where you tell it which class to look for and "trigger" to store on. Is it triggering and storing the results for both classes? Or just the first class? I would prefer it to store only the class of interest. Based on the saved wav files im getting, it seems to be storing both, but I wanted to ask because that could just be my model performing poorly.

I was also confused by the sleep record cycle setting. The default is 30 sleep duration, 30 record duration. Is this seconds or minutes?

Additonally, I was confused by the storage of the wav files. It seems like it combines the 'triggered' clips into 1 wav file? Is this the intended outcome?

Very cool tool I just want to get some clarity about usage. Thanks!