While training the model, are you converting the image in some range value or directly using the value from (0, 255)?
While training the model, are you converting the image in some range value or directly using the value from (0, 255)?