step-02.04-LogReg-with-Tensorflow.md

Logistic Regression with TensorFlow [back]

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)

1. You can Run TensorFlow Logistic Regression example from command-line:

$ cd ~/gitlab.altoros/776_DL_Libs_Benchmark.git/src/Step02/SubStep-04-Tensorflow
$ python ./TensorFlow_LogisticRegression.py

or with jupyter notebook:

$ cd ~/gitlab.altoros/776_DL_Libs_Benchmark.git/src/Step02/SubStep-04-Tensorflow
$ jupyter notebook TensorFlow_LogisticRegression_Notebook.ipynb

and then open in browser URL: http://ec2-54-86-145-119.compute-1.amazonaws.com:9999

---

Accuracy plot:

Network weights visualization: