Image Captioning

There is one variable missing in the second last video. I am unable to retrieve its value. Even if I provide custom value of steps_per_epoch, the algorithm is not training. I would really appreciate some kind assistance.