Getting issue of Classification metrics can't handle a mix of binary and unknown targets

I am trying to do classification of Cracks. I have uploaded images of positive and negative lablels. Created Edge impulse. But while doing transfer learning, i am facing below issue.
“/app/keras/.venv/lib/python3.8/site-packages/sklearn/metrics/”, line 93, in _check_targets raise ValueError( ValueError: Classification metrics can’t handle a mix of binary and unknown targets Model training complete.

I have tried inferencing on RPi 3b. It’s running, but its not classifying positive and negative images. Would you please help me out?

What change i should do to do classification properly?


Make you project public so I can see the data you labeled. My guess is you have some Samples that are not labeled. If the Samples are not cracks (like you have a photo of a tree) the label it unknown.