How can I visualize the architecture of a neural network using Keras?

Question

I tried the following:

#!/usr/bin/env python import keras from keras.models import model_from_yaml model_file_path = 'model-301.yaml' weights_file_path = 'model-301.hdf5' # Load network with open(model_file_path) as f: yaml_string = f.read() model = model_from_yaml(yaml_string) model.load_weights(weights_file_path) model.compile(optimizer='adagrad', loss='binary_crossentropy') # Visualize from keras.utils.visualize_util import plot 

However, this gives an error:

 Traceback (most recent call last): File "<stdin>", line 1, in <module> File "/home/moose/.local/lib/python2.7/site-packages/keras/utils/visualize_util.py", line 7, in <module> if not pydot.find_graphviz(): AttributeError: 'module' object has no attribute 'find_graphviz' 

How can i fix this?

Note. The hdf5 and YAML file can be found on Github .

Answer 1

The issue is also mentioned on the keras project issues page. You need to install pydot version <= 1.1.0, because the find_graphviz function was removed in version 1.2.0. Alternatively, you can install pydot-ng instead of recommended by keras developers.

Answer 2

If you have not installed the pydot python package pydot - try installing it. If you have pydot , reinstallation should help with your problem.