inconsistency in predicted probabilities for the same input

[01PrathamS]

i follow the full course, make an deep leanring model using InceptionV3 by following transfer learning but i want the model to produce the same probabilities consistently for repeated predictions on the same input.

scene1: i give input as sunflower and it predicts it right as 91% accuracy.

scene2: when i give same photo as input it predict is as 75% sugarcane

or sometimes it missclassify it also ..... i think this is not feasible.

please mark your guide, any advice or resources to be share is appreciatable. happy hacking