Transfer_learning_with_MobileNet_v1 Confusion


I have just started the second assignment for week 2 and I am finding the instructions for the first graded function to be unclear. It appears to me that we are to use tf.keras.Sequential([,]) to employ the flip and rotation, similar to how it is shown in the tf docs.

However, I am unsure what the point of the data_augmentation.add(None) is. I am pretty sure I am defining data_augmentation correctly but I just don’t know what is being attempted. Could someone explain the intent to this function so I can take another crack at figuring out the code?

Where a line of code in the notebook has “None”, it generally means that you’re supposed to replace “None” with some valid code for the task at hand.

There are two ways to specify a Sequential model.

  • One, the Sequential model asks you to define the layers you want to use as a list, separated by commas, inside a set of square brackets.
  • Two, you define a Sequential model starting with a blank list, and then you use the .add method to add new layers to it.

A lot of the exercises assume you’re already pretty fluent in both python and TensorFlow.

1 Like

Thank you kindly! I’m pretty sure I can figure it out from here now.

I am solid in R, a beginner in python, and my only exposure to tensor flow has been the course… Hopefully I’m not in over my head.

If you are enrolled, you can go back to Course 2 and do the TensorFlow exercises there.

Thanks @TMosh for the tips! I had used the [,] approach and just needed to comment out the .add lines.

I will definitely go back to course two and polish my tensorflow.