I would be extremely proud to get this Certificate even if it was useless to me in terms of ROI. But I have noticed so far in this course one deprecated Tensorflow module used in the Lab 1 and a Python module that is not only deprecated but has been removed from version 3.11 of Python. Certainly this is not good with regard to its negative reflection on Deeplearning.ai, but it has potential to cause negative effects on learning. The certification most definitely does not represent cutting-edge know-how.
Could you please be specific about your concerns?
On the other hand, do remember to use the recommended tensorflow APIs for data preprocessing and transformations. The staff are aware about this as well.
from tensorflow.keras.preprocessing.image import ImageDataGenerator
Module: tf.keras.preprocessing ← this is greyed out in v2.13 and there was an amber underline in the colab import statement.
I guess I got the wrong impression. If you are telling me you are on top of it, I am relieved.
Deprecated since version 3.9, removed in version 3.11: The optional parameter random.
Just IMHO the sample method seems like a better choice.
You are correct about
tf.keras.preprocessing being deprecated in tensorflow 2.13.x. Please use the suggested APIs by tensorflow. The staff are aware of this issue and are likely to update the labs as part of the next course update.
Please click my name and message your notebook as an attachment. I’d like to see where
random.shuffle is used.
I am wrong. I have been having some brain fog and trying to move through it. My claim that random.shuffle is suggested is of course nonsense. Again, I am sorry.
No worries. Thanks for confirming, Doug. Happy learning.