Python Keras module & #39; keras. backend & #39; has no attribute & #39; image_data_format & #39;, keraskeras. backend

Python Keras module 'keras. backend' has no attribute 'image _ data_format ', keraskeras. backend
Problem:

When the sample program mnist_cnn is run using Keras, the following error occurs: 'keras. backend' has no attribute 'image _ data_format'

Program path https://github.com/fchollet/keras/blob/master/examples/mnist_cnn.py

The python conda environment used is the carnd-term1 of the udacity self-driving course

 

Fault Section:

if K.image_data_format() == 'channels_first':    x_train = x_train.reshape(x_train.shape[0], 1, img_rows, img_cols)    x_test = x_test.reshape(x_test.shape[0], 1, img_rows, img_cols)    input_shape = (1, img_rows, img_cols)else:    x_train = x_train.reshape(x_train.shape[0], img_rows, img_cols, 1)    x_test = x_test.reshape(x_test.shape[0], img_rows, img_cols, 1)    input_shape = (img_rows, img_cols, 1)

Complete code see https://github.com/fchollet/keras/blob/master/examples/mnist_cnn.py

 

Fault analysis:

The Keras version ratio subprogram in the conda environment has an old version, so there is no 'image _ data_format 'variable.

 

Solution:

Either of the following two methods

1) if you do not upgrade the Keras version

Set K. image_data_format () = 'channels _ first'ReplaceK.image_dim_ordering() == 'th'

 

2) upgrade Keras to the latest version.

> Activate carnd-term1 // activate your conda environment, which is called carnd-term1

(Carnd-term1)> conda list // display packages installed in the current environment

(Carnd-term1)> pip uninstall Keras // uninstall the older version of Keras, which is installed using pip, so both uninstall and reinstall use pip-related commands

(Carnd-term1)> pip install Keras // reinstall the new version of Keras

(Carnd-term1)> conda list // check if version is updated