How many images should I train if I am applying i augmentation

#1

Suppose I have 800 images in my training data. Now I apply image augmentation. Image augmentation results in infinite images.

  1. How many images should I now be training?

  2. The number of images I train will depend upon steps per epoch?

#2

  1. it is upto you to decide how many images you want to train on in a single epoch but usually people prefer to train on number of images equal to the size of dataset, thats why step_per_epoch is set to len(train_set)/batch_size

  2. yes steps per epoch decides how many images get trained in a epoch more precisely it is steps_per_epoch multiplied by batch_size