Visualizing Model Structures in Keras

Update 3/May/2017: The steps mentioned in this post need to be slightly changed with the updates in Keras v2.*. Please check the updated guide here: Visualizing Keras Models - Updated.

Have you ever wanted to visualize the structure of a Keras model? When you have a complex model, sometimes it's easy to wrap your head around it if you can see a visual representation of it. What if there's a way to automatically build such a visual representation of a model?

Well, there is a way. Keras has a model visualization function, that can plot out the structure of a model. It would look something like this,

The visualization of the LeNet model

Above is the visualization of the LeNet model, which is defined in code as follows,

 # initialize the model  
 model = Sequential()  
   
 # first set of CONV => RELU => POOL  
 model.add(Convolution2D(20, 5, 5, border_mode="same",  
     input_shape=(height, width, depth)))  
 model.add(Activation("relu"))  
 model.add(MaxPooling2D(pool_size=(2, 2), strides=(2, 2)))  
   
 # second set of CONV => RELU => POOL  
 model.add(Convolution2D(50, 5, 5, border_mode="same"))  
 model.add(Activation("relu"))  
 model.add(MaxPooling2D(pool_size=(2, 2), strides=(2, 2)))  
   
 # set of FC => RELU layers  
 model.add(Flatten())  
 model.add(Dense(500))  
 model.add(Activation("relu"))  
   
 # softmax classifier  
 model.add(Dense(classes))  
 model.add(Activation("softmax"))  

The documentation page of the Keras Visualizer states that you just need 2 lines of code to plot a model,

 from keras.utils.visualize_util import plot  
 plot(model, to_file='model.png')  

Sounds easy right? But, as soon as you try to run, you learn that you need to setup few dependencies in order for it to run. Basically, you need to,

1. Install graphviz Python package, and the graphviz binaries added to the PATH.
2. Install pydot Python package
3. Fix a bug in the Visualization module in Keras

The graphviz conda package only contains the binaries, and not the Python bindings. The pip package for graphviz only has the Python bindings and not the binaries. So, we need to install both.
First, install conda package,

 conda install graphviz  

This places the graphviz binary files in to the <path to anaconda environment>\Library\bin\graphviz\ directory. So, you need to add it to the system PATH.
Then, we install the graphviz pip package,

 pip install graphviz  

Then, we install pydot from pip,

 pip install pydot  

Then, we need to do an edit in the Keras Visualization module. There is a bug in that code, which doesn't work with the latest version of pydot.
Open the <path to anaconda environment>\lib\site-packages\keras\utils\visualize_util.py file, and comment out the following block,

 #if not pydot.find_graphviz():  
 #  raise ImportError('Failed to import pydot. You must install pydot'  
 #           ' and graphviz for `pydotprint` to work.')  

The pydot.find_graphviz() function does not exist in the latest version of pydot.

After these steps are done, you should be able to import the visualize_util and run the plot function to generate the visualization.

Try fiddling with the show_shapes and show_layer_names parameters in the plot function to see how it changes the generated graph.

Related links:
https://keras.io/visualization/

Build Deeper: Deep Learning Beginners' Guide is the ultimate guide for anyone taking their first step into Deep Learning.

Get your copy now!