tf-rnn-nlg

This is a simple RNN that builds a small essay based on its input data. To update the data, change the text in data/input.txt or specify an argument.

For convenience, I have an already trained model created.

Steps to train a new model

• run 'python train.py' and it will train a new model based on the input data
• Please note that this will take a while, since this is currently CPU Bound

Steps to generate a new essay.

• run 'python sample.py' and it will generate a new essay based on the model trained in save/
• The result will be saved to save/result.txt