Home/ Questions/Attributeerror: module 'tensorflow' has no attribute 'get_default_graph' - how can I circumvent the error?
Next
Answered
Romy Wang
  • 22
Asked: 2022-05-18T20:53:29+00:00 In:

Attributeerror: module ‘tensorflow’ has no attribute ‘get_default_graph’ – how can I circumvent the error?

  • 22

. Advertisement .

..3..

. Advertisement .

..4..

For the problem “attributeerror: module ‘tensorflow’ has no attribute ‘get_default_graph’.” I tried to fix it, but it doesn’t work and returns the result I want. Here is my program:

from keras.models import Sequential
 
 model = Sequential()
 model.add(Dense(32, input_dim=784))
 model.add(Activation('relu'))
 model.add(LSTM(17))
 model.add(Dense(1, activation='sigmoid'))
 model.compile(loss='binary_crossentropy', optimizer='adam', metrics=['accuracy'])

and

AttributeError: module 'tensorflow' has no attribute 'get_default_graph'

has occurred. I’ve checked the entire command line but still can’t find the mistake.

module tensorflow has no attribute get_default_graph
Share

2 Answers

  1. Best Answer
    Expert

    The cause: Tensorflow version changes are the cause of this attributeerror: module ‘tensorflow’ has no attribute ‘get_default_graph’ error.

    Solution:

    Change:

    from keras.models import Sequential

    into:

    from tensorflow.keras.models import Sequential
    • 0

  2. tf.compat.v1.get_default_graph() was used for tf 2.1.0 – e.g. 

    import tensorflow as tf
sess = tf.compat.v1.Session(graph=tf.compat.v1.get_default_graph(), config=session_conf)
tf.compat.v1.keras.backend.set_session(sess)
    • 20

Leave an answer

You must login to add an answer.

Need An Account,