This is the 3rd project for Udacity Deep Learning nanodegree. The aim of the project is to build a recurrent neuron network (RNN) and generate a script based on the model.
The repository includes:
the python notebook file for training the data and generate the script using the trained model. (Adapted from Udacity DL Nanodegree Project: Generate a TV script.); a downloaded html file to show the results; a saved model to generate the script; other files for deployment and supportive files
You can load the model to generate your own result without training a new one ab initio.
The parameters for training the model are:
num_epochs = 10 learning_rate = 0.001 vocab_size = len(vocab_to_int) output_size = vocab_size embedding_dim = 256 hidden_dim = 512 n_layers = 3 show_every_n_batches = 500