Overview

This repository contains simple examples to understand TensorFlow from the ground up. One dataset contains advertising expenditure across various channels and the sales achieved. This dataset is used in most examples.

The following two examples are shown:

• Linear Regression: Example lin_reg_1 solves the problem from first principles, based on the solution to a differential equation that minimizes the cost function. Example lin_reg_2 gives the solution using a Machine Learning approach with the use of TensorFlow.
• Neural Networks: Example nn_python solves the problem without using TensorFlow, with nothing more than Python, NumPy and Pandas. Example nn_tf gives the solution using a Machine Learning approach with the use of TensorFlow.

Example nn_example uses the famous MNIST data. For brevity, MNIST data is not included in this repository. Download all four *.gz files from the above link and place them in data/ folder.

Refer to docs/ folder for some background information to understand this code.