Last answered:

15 Jun 2020

Posted on:

21 May 2020

0

Tensorflow

What is the difference between the two courses Tensorflow and Tensorflow 2.0?

i also have a problem initial to run tensorflow but after using web suggestions:

import tensorflow.compat.v1 as tf
tf.disable_v2_behavior()

it is solved. Is this a version problem?

i try with the installation guide in tensorflow 2.0.

i could not find the created enviroment through kernel in jupyter. but i follow and print version, i have tensor 2.2.0

however i still encounter this error:

ModuleNotFoundError                       Traceback (most recent call last)
<ipython-input-1-9c8e25c07922> in <module>
      8 # conda install tensorflow-datasets
      9 
---> 10 import tensorflow_datasets as tfds
     11 
     12 # these datasets will be stored in C:\Users\*USERNAME*\tensorflow_datasets\...

ModuleNotFoundError: No module named 'tensorflow_datasets'


Now I try:

from tensorflow.keras.datasets import mnist

this error surface:

AttributeError                            Traceback (most recent call last)
<ipython-input-4-8366146aadf2> in <module>
     20 # Have in mind that the very first time you execute this command it might take a little while to run
     21 # because it has to download the whole dataset. Following commands only extract it so they're faster.
---> 22 mnist = mnist.read_data_sets("MNIST_data/", one_hot=True)

AttributeError: module 'tensorflow.keras.datasets.mnist' has no attribute 'read_data_sets'
 
3 weeks, no answer, but i found answer in other's questions. thanks so much.
1 answers ( 0 marked as helpful)
Instructor
Posted on:

15 Jun 2020

0
Hi Kelvin, Thanks for reaching out.  First, you can check this video for more information on the differences between TF1 and TF2: https://www.youtube.com/watch?v=t48a_KOh0fQ Regarding the environment - if you cannot see it in Anaconda, then you have not created it. Regarding the tensorflow_datasets issue -> please install the library using: pip install tensorflow_datasets What is the version of your TensorFlow? Please check that with: print(tf.__version__) Best, The 365 Team

Submit an answer