kahst/BirdNET-Analyzer

Training data not specific enough

LClelland opened this issue · 2 comments

I have made a model to detect wildcat vocalisations and have created a tflite file, however when I try to test it it decides that everything is a wildcat. This was true using a file that does have wildcat vocalisations and one that does not. Is there any way to make the model more specific?

What were your training classes?