step-02.01.03-Theano-Keras-Framework.md

Logistic Regression with Theano and Keras Framework [back]

At this point Theano is already installed in the system.

1. Install Keras:

$ sudo pip install keras

2. Install addition Python-SVG modules:

$ sudo apt-get install python-pydot
$ sudo pip install pydot -U

3. To run example code:

$ cd ~/gitlab.altoros/776_DL_Libs_Benchmark.git/src/Step02/SubStep-01.03-Theano-Keras
$ jupyter notebook Theano_Keras_LogisticRegression_Notebook.ipynb

and go to the URL: http://ec2-54-172-161-206.compute-1.amazonaws.com:9999/

3. To work with the MNIST dataset we will use a helper method from Tensorflow package:

from tensorflow.examples.tutorials.mnist import input_data
mnist = input_data.read_data_sets("MNIST_data/", one_hot=False)

---

Network visualisation:

---

Classification visualisation:

---

Weights visualisation: