The problem Plant Disease Diagnosis Using Deep Learning solves

The Application is developed to classify images in offline mode , saving a need for server architecture .
TFLite model used for classification has an accuracy of 97%

Challenges I ran into

Conversion of Pytorch to TFLite model : Solution came out was to convert PyTorch model to Keras then to TFlite .
Image preprocessing at android Side , before feeding the images to the model .