A Companion to TensorFlow Udacity MOOC (WIP)

This repository aims at accompanying and complementing the Python code of the TensorFlow Udacity MOOC by Google. It provides elaborate solutions to the MOOC assignments. When deemed useful, alternative solutions or elaborations are provided to clarify concepts or methods.

Jupyter notebooks are used for formatting the code in order to display textual and visual output along with code. The notebooks have been developed using Python 3.6.6 (WIP). If you would like to recommend any improvements, feel free to open a PR or issue.

Before getting started, two steps need to be taken:

1. Every notebook sets BASEDIR. You need to set BASEDIR to the root of your cloned git repository.
2. For assignment 1, download the notMNIST data. More specifically, download the large notMNIST and small notMNIST datasets, decompress them and place them in BASEDIR/data/notmnist/original/notmnist_large and BASEDIR/data/notmnist/original/notmnist_small, respectively. The large and small notMNIST datasets correspond to the training and test set. For instance, the training data for letter A is located in BASEDIR/data/notmnist/original/notmnist_large/A after downloading, decompressing and moving the training data.