为什么 tf.keras 不起作用，但 tensorflow.keras 是否有效？

Question

I've used the following lines of code:

!pip install --upgrade tensorflow
import tensorflow as tf

from tensorflow.keras import models - this works fine

BUT

When from tf.keras import models is used, error appears

---------------------------------------------------------------------------
ModuleNotFoundError                       Traceback (most recent call last)
<ipython-input-5-15abaed607f8> in <module>()
----> 1 from tf.keras import models

ModuleNotFoundError: No module named 'tf'

---------------------------------------------------------------------------
NOTE: If your import is failing due to a missing package, you can
manually install dependencies using either !pip or !apt.

To view examples of installing some common dependencies, click the
"Open Examples" button below.
---------------------------------------------------------------------------

Why is this happening?

Answer 1

Packages are stored in the computers as folders. For example, there is a xgboost folder if you import that. When you type

from tf.keras import models

Python will look for a 'tf' reference (package) in your system, which does not exist, tf just exist as an alias in your python IDE.