Tensorflow not updated to later version

Course Q&A Deep Learning Specialization Convolutional Neural Networks
,
1

Hi,

I’m working on Lab 2 of Week 2 in the Convolutional Neural Networks course. In the first graded function (data_augmenter), I’m required to use tf.keras.layers.RandomFlip and RandomRotation as part of the data augmentation pipeline.

However, the lab environment is running TensorFlow 2.3, and these layers were only introduced in TensorFlow 2.4 and above. As a result, I get an AttributeError stating that RandomFlip does not exist.

Since the data_augmenter function is used throughout the rest of the lab, I cannot proceed or pass the grader.

I need to know what I can do about it.

Thank you!

2

They specifically imported those functions for you in the “import” cell at the beginning of the notebook. They had to use a different superclass than tf.keras.layers because those functions were still considered “experimental” in the version of TF we are using here.

1 Like