ImportError: No module named 'tensorflow.python'

Question

here i wanna run this code for try neural network with python :

from __future__ import print_function 
from keras.datasets import mnist from 
keras.models import Sequential from 
keras.layers import Activation, Dense 
from keras.utils import np_utils 
import tensorflow as tf


batch_size = 128 nb_classes = 10 nb_epoch = 12

#input image dimensions img_row, img_cols = 28, 28

#the data, Shuffled and split between train and test sets (X_train, y_train), (X_test, y_test) = mnist.load_data()


X_train = X_train.reshape(X_train.shape[0], img_rows * img_cols)

X_test = X_test.reshape(X_test.shape[0], img_row * img_cols)

X_train = X_train.astype('float32') X_test = X_test.astype('float32') X_train /= 255 X_text /= 255

print('X_train shape:', X_train.shape) print(X_train_shape[0], 'train samples') print(X_test_shape[0], 'test samples')

#convert class vectors to binary category

Y_train = np_utils.to_categorical(y_train, nb_classes)

Y_test = np_utils.to_categorical(y_test, nb_classes)

model = Sequential()

model.add(Dense(output_dim = 800, input_dim=X_train.shape[1])) model.add(Activation('sigmoid')) model.add(Dense(nb_classes)) model.add(Actiovation('softmax'))

model.compile(loss = 'categorical_crossentropy', optimizer='sgd', metrics=['accuracy']) #crossentropy fungsi galat atau fungsi error dipakai kalo class biner



#model.fit(X_train, Y_train, batch_size=batch_size, nb_epoch = nb_poch, verbose=1, validation_data=(X_test, Y_test))

score = model.evaluate(X_test, Y_test, verbose = 0) print('Test Score : ', score[0]) print('Test Accuracy : ', score[1])

at the beginning it must install keras, and success. but when try to run the code at the first the error is :

ImportError : No Moduled Name "tensorflow"

then i install using pip :

pip install tensorflow

after installation i try to run code again, got another message like this :

ImportError : No Moduled Name "tensorflow.python"

Message Error i dont have any idea with the error

Solution 1

Uninstall tensorflow:

pip uninstall tensorflow

Then reinstall it:

pip install tensorflow

Solution 2

for me upgrading pip helped,

pip install --upgrade pip
pip uninstall tensorflow
pip install tensorflow

Solution 3

If you have python 3.6 and up (most likely), a pip3 package will be installed by default. Installing tensorflow using pip3 will make the path of the installation visible to python. So try

pip3 install tensorflow

First response ever, hope it helps!

Solution 4

I have the same problem in Windows 10. Until now I don't know why.

But if I create an virtual environment

cd <your project path>

Install virtualenv

pip install virtualenv

Create the virtual environment

virtualenv <envname>

Activate the env

• Windows Powershell: .\<envname>\Scripts\activate
• Unix with Bash or zsh: source <envname>/bin/activate

Then now you install tensorflow

(<envname>) $ pip install tensorflow

And then run Hello World successfully.

*Don't forget that you need to activate or configure everytime the virtual environment jupyter, command-line, etc.

View more solutions

Comments

[ppasler]

Is it this one: github.com/tensorflow/tensorflow/issues/2885

[Boooooooooms]

My tensorflow is in help('modules') list.

[Oshada]

executing sys.path() gives this error TypeError: 'list' object is not callable I think it should be sys.path as it is not a method

[Sumax]

Hi @zhaozheng would you be able to explain as to why this occurs; in the original tensorflow, why is the tensorflow.python not properly installed with it.

[Sumant Agnihotri]

Not sure if I should be happy or sad that this actually worked

[Karan]

same with me , this method even worked for me even though my error was different :)

[Matthew Barlowe]

The OP mentions in the question they have installed tensor flow with pip this answer is not helpful

[mjkvaak]

If the above does not help, you can try installing non-cached tensorflow by adding --no-cache-dir to the last code line.

[Stack crashed]

This 'works' because you most likely had a gpu based tensorflow installed before. By uninstalling and reinstalling you just changed that to cpu. Since there are other dependencies for gpu support, it had issues, now you don't face those issues because it is "downgraded" to cpu.