recurrent-neural-nets-tensorflow

Implementation of Recurrent Neural Nets in Tensorflow

The aim of the project is to implement the forward pass from scratch for a Recurrent Neural Net using Tensorflow.

The implementation covers the following points:

1. Workflow for building and using the computational graph in Tensorflow
2. Recurrent NN implementation and application
3. Training word embeddings as an additional layer in the front of the NN

The project implements a simple Part of Speech Tagging netword using CoNLL 2000 Chunking data (http://www.cnts.ua.ac.be/conll2000/chunking/)

Resources

1. http://www.wildml.com/2015/09/recurrent-neural-networks-tutorial-part-1-introduction-to-rnns/