how to adapt output of a keras model
I want to create a fruit ripness detector.
For that I could use the API of roboflow but I want to generate a trained model locally.
I downloaded the dataset from this page: https://universe.roboflow.com/mixed-fruit-annotation/fruit-ripness-detector/dataset/2/download/tfrecord
Machine Learning Visual model with 95% accuracy predicts new data the same label
[confusion matrix (https://i.sstatic.net/WiTyUOew.png)]