BirdNET v2.5 model species classes?

[tphakala]

According to this video (https://www.youtube.com/watch?v=Faavvmi9JZw), the upcoming v2.5 model will add a significant number of non-bird sounds to its detection library. Currently, BirdNET already detects some non-bird species, but there's no clear classification, in labels or any external resource, to indicate which family or category each detected sound belongs to (e.g., bird, amphibian, insect, canine, man-made sound, etc.).

Could the v2.5 model address this issue?

[kahst]

We'll use iNatSounds data to add non-birds and what they do is add the taxonomic group to the label name. I think this is what we'll do too which should help to distinguish birds from non-birds. We'll also try to update the geomodel to include the non-birds.