TV Script Generation

Student: Angel Martinez-Tenor
Deep Learning Nanodegree Foundation - Udacity
March 26, 2017

In this project, you'll generate your own Simpsons TV scripts using RNNs. You'll be using part of the Simpsons dataset of scripts from 27 seasons. The Neural Network you'll build will generate a new TV script for a scene at Moe's Tavern.

Get the Data

The data is already provided for you. You'll be using a subset of the original dataset. It consists of only the scenes in Moe's Tavern. This doesn't include other versions of the tavern, like "Moe's Cavern", "Flaming Moe's", "Uncle Moe's Family Feed-Bag", etc..


In [1]:
"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
import helper

data_dir = './data/simpsons/moes_tavern_lines.txt'
text = helper.load_data(data_dir)
# Ignore notice, since we don't use it for analysing the data
text = text[81:]

Explore the Data

Play around with view_sentence_range to view different parts of the data.


In [2]:
view_sentence_range = (0, 10)

"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
import numpy as np

print('Dataset Stats')
print('Roughly the number of unique words: {}'.format(len({word: None for word in text.split()})))
scenes = text.split('\n\n')
print('Number of scenes: {}'.format(len(scenes)))
sentence_count_scene = [scene.count('\n') for scene in scenes]
print('Average number of sentences in each scene: {}'.format(np.average(sentence_count_scene)))

sentences = [sentence for scene in scenes for sentence in scene.split('\n')]
print('Number of lines: {}'.format(len(sentences)))
word_count_sentence = [len(sentence.split()) for sentence in sentences]
print('Average number of words in each line: {}'.format(np.average(word_count_sentence)))

print()
print('The sentences {} to {}:'.format(*view_sentence_range))
print('\n'.join(text.split('\n')[view_sentence_range[0]:view_sentence_range[1]]))


Dataset Stats
Roughly the number of unique words: 11492
Number of scenes: 262
Average number of sentences in each scene: 15.248091603053435
Number of lines: 4257
Average number of words in each line: 11.50434578341555

The sentences 0 to 10:
Moe_Szyslak: (INTO PHONE) Moe's Tavern. Where the elite meet to drink.
Bart_Simpson: Eh, yeah, hello, is Mike there? Last name, Rotch.
Moe_Szyslak: (INTO PHONE) Hold on, I'll check. (TO BARFLIES) Mike Rotch. Mike Rotch. Hey, has anybody seen Mike Rotch, lately?
Moe_Szyslak: (INTO PHONE) Listen you little puke. One of these days I'm gonna catch you, and I'm gonna carve my name on your back with an ice pick.
Moe_Szyslak: What's the matter Homer? You're not your normal effervescent self.
Homer_Simpson: I got my problems, Moe. Give me another one.
Moe_Szyslak: Homer, hey, you should not drink to forget your problems.
Barney_Gumble: Yeah, you should only drink to enhance your social skills.


Implement Preprocessing Functions

The first thing to do to any dataset is preprocessing. Implement the following preprocessing functions below:

  • Lookup Table
  • Tokenize Punctuation

Lookup Table

To create a word embedding, you first need to transform the words to ids. In this function, create two dictionaries:

  • Dictionary to go from the words to an id, we'll call vocab_to_int
  • Dictionary to go from the id to word, we'll call int_to_vocab

Return these dictionaries in the following tuple (vocab_to_int, int_to_vocab)


In [3]:
import numpy as np
import problem_unittests as tests

def create_lookup_tables(text):
    """
    Create lookup tables for vocabulary
    :param text: The text of tv scripts split into words
    :return: A tuple of dicts (vocab_to_int, int_to_vocab)
    """
    # TODO: Implement Function
    
    int_to_vocab = dict(enumerate(set(text)))  #  direct dict from list
    vocab_to_int = {word: i for i, word in int_to_vocab.items()}  # reverse dict
    
    return vocab_to_int, int_to_vocab
   
    
"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_create_lookup_tables(create_lookup_tables)


Tests Passed

Tokenize Punctuation

We'll be splitting the script into a word array using spaces as delimiters. However, punctuations like periods and exclamation marks make it hard for the neural network to distinguish between the word "bye" and "bye!".

Implement the function token_lookup to return a dict that will be used to tokenize symbols like "!" into "||Exclamation_Mark||". Create a dictionary for the following symbols where the symbol is the key and value is the token:

  • Period ( . )
  • Comma ( , )
  • Quotation Mark ( " )
  • Semicolon ( ; )
  • Exclamation mark ( ! )
  • Question mark ( ? )
  • Left Parentheses ( ( )
  • Right Parentheses ( ) )
  • Dash ( -- )
  • Return ( \n )

This dictionary will be used to token the symbols and add the delimiter (space) around it. This separates the symbols as it's own word, making it easier for the neural network to predict on the next word. Make sure you don't use a token that could be confused as a word. Instead of using the token "dash", try using something like "||dash||".


In [4]:
def token_lookup():
    """
    Generate a dict to turn punctuation into a token.
    :return: Tokenize dictionary where the key is the punctuation and the value is the token
    """
    # TODO: Implement Function
    
    punc = ('.', ',', '"', ';', '!', '?', '(', ')', '--', '\n')
    token = ('Period', 'Comma', 'Quotation_Mark', 'Semicolon', 'Exclamation_Mark', 
             'Question_Mark', 'Left_Parentheses', 'Right_Parentheses', 'Dash', 'Return')   
    
    punc_to_token = {p: "||"+t+"||" for p,t in zip(punc,token)}

    return punc_to_token

"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_tokenize(token_lookup)


Tests Passed

Preprocess all the data and save it

Running the code cell below will preprocess all the data and save it to file.


In [5]:
"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
# Preprocess Training, Validation, and Testing Data
helper.preprocess_and_save_data(data_dir, token_lookup, create_lookup_tables)

Check Point

This is your first checkpoint. If you ever decide to come back to this notebook or have to restart the notebook, you can start from here. The preprocessed data has been saved to disk.


In [6]:
"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
import helper
import numpy as np
import problem_unittests as tests

int_text, vocab_to_int, int_to_vocab, token_dict = helper.load_preprocess()

Build the Neural Network

You'll build the components necessary to build a RNN by implementing the following functions below:

  • get_inputs
  • get_init_cell
  • get_embed
  • build_rnn
  • build_nn
  • get_batches

Check the Version of TensorFlow and Access to GPU


In [7]:
"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
from distutils.version import LooseVersion
import warnings
import tensorflow as tf

# Check TensorFlow Version
assert LooseVersion(tf.__version__) >= LooseVersion('1.0'), 'Please use TensorFlow version 1.0 or newer'
print('TensorFlow Version: {}'.format(tf.__version__))

# Check for a GPU
if not tf.test.gpu_device_name():
    warnings.warn('No GPU found. Please use a GPU to train your neural network.')
else:
    print('Default GPU Device: {}'.format(tf.test.gpu_device_name()))


TensorFlow Version: 1.0.1
Default GPU Device: /gpu:0

Input

Implement the get_inputs() function to create TF Placeholders for the Neural Network. It should create the following placeholders:

  • Input text placeholder named "input" using the TF Placeholder name parameter.
  • Targets placeholder
  • Learning Rate placeholder

Return the placeholders in the following the tuple (Input, Targets, LearingRate)


In [8]:
def get_inputs():
    """
    Create TF Placeholders for input, targets, and learning rate.
    :return: Tuple (input, targets, learning rate)
    """
    # TODO: Implement Function
    
    input_text = tf.placeholder(tf.int32, [None, None], name='input')
    targets = tf.placeholder(tf.int32, [None, None])
    learning_rate = tf.placeholder(tf.float32)
    
    return input_text, targets, learning_rate


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_get_inputs(get_inputs)


Tests Passed

Build RNN Cell and Initialize

Stack one or more BasicLSTMCells in a MultiRNNCell.

  • The Rnn size should be set using rnn_size
  • Initalize Cell State using the MultiRNNCell's zero_state() function
    • Apply the name "initial_state" to the initial state using tf.identity()

Return the cell and initial state in the following tuple (Cell, InitialState)


In [9]:
def get_init_cell(batch_size, rnn_size):
    """
    Create an RNN Cell and initialize it.
    :param batch_size: Size of batches
    :param rnn_size: Size of RNNs
    :return: Tuple (cell, initialize state)
    """
    # TODO: Implement Function
    
    lstm_layers = 1  # Two lstm layers improves the trail_loss by 0.1 but takes x4 time
    keep_prob = 0.8  # adding a 20% dropout layer improved the training
    
    rnn = tf.contrib.rnn.BasicLSTMCell(rnn_size)
    drop = tf.contrib.rnn.DropoutWrapper(rnn, output_keep_prob=keep_prob)
    cell = tf.contrib.rnn.MultiRNNCell([drop] * lstm_layers)
    initial_state = tf.identity(cell.zero_state(batch_size, tf.float32), 'initial_state')
    
    return cell, initial_state


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_get_init_cell(get_init_cell)


Tests Passed

Word Embedding

Apply embedding to input_data using TensorFlow. Return the embedded sequence.


In [10]:
def get_embed(input_data, vocab_size, embed_dim):
    """
    Create embedding for <input_data>.
    :param input_data: TF placeholder for text input.
    :param vocab_size: Number of words in vocabulary.
    :param embed_dim: Number of embedding dimensions
    :return: Embedded input.
    """
    # TODO: Implement Function
    
    embedding = tf.Variable(tf.random_uniform([vocab_size, embed_dim], -1.0, 1.0))
    embedded_input = tf.nn.embedding_lookup(embedding, input_data)
    
    return embedded_input


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_get_embed(get_embed)


Tests Passed

Build RNN

You created a RNN Cell in the get_init_cell() function. Time to use the cell to create a RNN.

Return the outputs and final_state state in the following tuple (Outputs, FinalState)


In [11]:
def build_rnn(cell, inputs):
    """
    Create a RNN using a RNN Cell
    :param cell: RNN Cell
    :param inputs: Input text data
    :return: Tuple (Outputs, Final State)
    """
    # TODO: Implement Function
    
    outputs, final_state = tf.nn.dynamic_rnn(cell, inputs, dtype=tf.float32)
    final_state = tf.identity(final_state, "final_state")
    
    return outputs, final_state


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_build_rnn(build_rnn)


Tests Passed

Build the Neural Network

Apply the functions you implemented above to:

  • Apply embedding to input_data using your get_embed(input_data, vocab_size, embed_dim) function.
  • Build RNN using cell and your build_rnn(cell, inputs) function.
  • Apply a fully connected layer with a linear activation and vocab_size as the number of outputs.

Return the logits and final state in the following tuple (Logits, FinalState)


In [12]:
def build_nn(cell, rnn_size, input_data, vocab_size):
    """
    Build part of the neural network
    :param cell: RNN cell
    :param rnn_size: Size of rnns
    :param input_data: Input data
    :param vocab_size: Vocabulary size
    :return: Tuple (Logits, FinalState)
    """
    # TODO: Implement Function
    
    # A hardcoded embed_dim can be also employed here, but rnn_size usually results in better performance
    
    rnn_inputs = get_embed(input_data, vocab_size, rnn_size)    
    outputs, final_state = build_rnn(cell, rnn_inputs)
    logits = tf.contrib.layers.fully_connected(outputs, vocab_size, activation_fn=None)
    
    return logits, final_state


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_build_nn(build_nn)


Tests Passed

Batches

Implement get_batches to create batches of input and targets using int_text. The batches should be a Numpy array with the shape (number of batches, 2, batch size, sequence length). Each batch contains two elements:

  • The first element is a single batch of input with the shape [batch size, sequence length]
  • The second element is a single batch of targets with the shape [batch size, sequence length]

If you can't fill the last batch with enough data, drop the last batch.

For exmple, get_batches([1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15], 2, 3) would return a Numpy array of the following:

[
  # First Batch
  [
    # Batch of Input
    [[ 1  2  3], [ 7  8  9]],
    # Batch of targets
    [[ 2  3  4], [ 8  9 10]]
  ],

  # Second Batch
  [
    # Batch of Input
    [[ 4  5  6], [10 11 12]],
    # Batch of targets
    [[ 5  6  7], [11 12 13]]
  ]
]

In [13]:
def get_batches(int_text, batch_size, seq_length):
    """
    Return batches of input and target
    :param int_text: Text with the words replaced by their ids
    :param batch_size: The size of batch
    :param seq_length: The length of sequence
    :return: Batches as a Numpy array
    """
    # TODO: Implement Function

    words_in_batch = batch_size*seq_length 
    
    n_batches = (len(int_text)-1)//words_in_batch # len-1 would avoid later error if int_text//batch_size == int_text/batch_size
    inputs = int_text[: n_batches*words_in_batch]
    targets = int_text[1 : n_batches*words_in_batch + 1]  
    
    # input and target sequences are defined here for clarity in the loop below
    input_sequences = [inputs[i:i+seq_length] for i  in range(0, len(inputs), seq_length)] 
    target_sequences = [targets[i:i+seq_length] for i  in range(0, len(targets), seq_length)] 

    batches = []
    
    for i in range(n_batches):
        input_batch = []
        target_batch = []
        for j in range(batch_size):
            input_batch.append(input_sequences[n_batches*j+i])
            target_batch.append(target_sequences[n_batches*j+i])
        batches.append((input_batch, target_batch))
    
    return np.array(batches)


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_get_batches(get_batches)


Tests Passed

Neural Network Training

Hyperparameters

Tune the following parameters:

  • Set num_epochs to the number of epochs.
  • Set batch_size to the batch size.
  • Set rnn_size to the size of the RNNs.
  • Set seq_length to the length of sequence.
  • Set learning_rate to the learning rate.
  • Set show_every_n_batches to the number of batches the neural network should print progress.

In [14]:
# Number of Epochs
num_epochs = 25   
# Batch Size
batch_size = 128  # Tested: from 64 to 512
# RNN Size
rnn_size = 1024  # Tested: from 128 to 1024
# Sequence Length
seq_length = 20  # Tested: from 10 to 30
# Learning Rate
learning_rate = 0.005 # Tested: from 0.0001 to 0.05
# Show stats for every n number of batches
show_every_n_batches = 10

"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
save_dir = './save'

Build the Graph

Build the graph using the neural network you implemented.


In [15]:
"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
from tensorflow.contrib import seq2seq

train_graph = tf.Graph()
with train_graph.as_default():
    vocab_size = len(int_to_vocab)
    input_text, targets, lr = get_inputs()
    input_data_shape = tf.shape(input_text)
    cell, initial_state = get_init_cell(input_data_shape[0], rnn_size)
    logits, final_state = build_nn(cell, rnn_size, input_text, vocab_size)

    # Probabilities for generating words
    probs = tf.nn.softmax(logits, name='probs')

    # Loss function
    cost = seq2seq.sequence_loss(
        logits,
        targets,
        tf.ones([input_data_shape[0], input_data_shape[1]]))

    # Optimizer
    optimizer = tf.train.AdamOptimizer(lr)

    # Gradient Clipping
    gradients = optimizer.compute_gradients(cost)
    capped_gradients = [(tf.clip_by_value(grad, -1., 1.), var) for grad, var in gradients]
    train_op = optimizer.apply_gradients(capped_gradients)

Train

Train the neural network on the preprocessed data. If you have a hard time getting a good loss, check the forms to see if anyone is having the same problem.


In [16]:
"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
batches = get_batches(int_text, batch_size, seq_length)

with tf.Session(graph=train_graph) as sess:
    sess.run(tf.global_variables_initializer())

    for epoch_i in range(num_epochs):
        state = sess.run(initial_state, {input_text: batches[0][0]})

        for batch_i, (x, y) in enumerate(batches):
            feed = {
                input_text: x,
                targets: y,
                initial_state: state,
                lr: learning_rate}
            train_loss, state, _ = sess.run([cost, final_state, train_op], feed)

            # Show every <show_every_n_batches> batches
            if (epoch_i * len(batches) + batch_i) % show_every_n_batches == 0:
                print('Epoch {:>3} Batch {:>4}/{}   train_loss = {:.3f}'.format(
                    epoch_i,
                    batch_i,
                    len(batches),
                    train_loss))

    # Save Model
    saver = tf.train.Saver()
    saver.save(sess, save_dir)
    print('Model Trained and Saved')


Epoch   0 Batch    0/26   train_loss = 8.821
Epoch   0 Batch   10/26   train_loss = 5.896
Epoch   0 Batch   20/26   train_loss = 5.411
Epoch   1 Batch    4/26   train_loss = 4.925
Epoch   1 Batch   14/26   train_loss = 4.640
Epoch   1 Batch   24/26   train_loss = 4.487
Epoch   2 Batch    8/26   train_loss = 4.290
Epoch   2 Batch   18/26   train_loss = 4.054
Epoch   3 Batch    2/26   train_loss = 3.692
Epoch   3 Batch   12/26   train_loss = 3.565
Epoch   3 Batch   22/26   train_loss = 3.309
Epoch   4 Batch    6/26   train_loss = 3.227
Epoch   4 Batch   16/26   train_loss = 3.117
Epoch   5 Batch    0/26   train_loss = 2.756
Epoch   5 Batch   10/26   train_loss = 2.788
Epoch   5 Batch   20/26   train_loss = 2.742
Epoch   6 Batch    4/26   train_loss = 2.460
Epoch   6 Batch   14/26   train_loss = 2.471
Epoch   6 Batch   24/26   train_loss = 2.312
Epoch   7 Batch    8/26   train_loss = 2.113
Epoch   7 Batch   18/26   train_loss = 2.109
Epoch   8 Batch    2/26   train_loss = 1.917
Epoch   8 Batch   12/26   train_loss = 1.717
Epoch   8 Batch   22/26   train_loss = 1.741
Epoch   9 Batch    6/26   train_loss = 1.598
Epoch   9 Batch   16/26   train_loss = 1.486
Epoch  10 Batch    0/26   train_loss = 1.363
Epoch  10 Batch   10/26   train_loss = 1.272
Epoch  10 Batch   20/26   train_loss = 1.169
Epoch  11 Batch    4/26   train_loss = 1.154
Epoch  11 Batch   14/26   train_loss = 1.091
Epoch  11 Batch   24/26   train_loss = 0.896
Epoch  12 Batch    8/26   train_loss = 0.861
Epoch  12 Batch   18/26   train_loss = 0.877
Epoch  13 Batch    2/26   train_loss = 0.811
Epoch  13 Batch   12/26   train_loss = 0.777
Epoch  13 Batch   22/26   train_loss = 0.748
Epoch  14 Batch    6/26   train_loss = 0.691
Epoch  14 Batch   16/26   train_loss = 0.678
Epoch  15 Batch    0/26   train_loss = 0.625
Epoch  15 Batch   10/26   train_loss = 0.570
Epoch  15 Batch   20/26   train_loss = 0.555
Epoch  16 Batch    4/26   train_loss = 0.547
Epoch  16 Batch   14/26   train_loss = 0.508
Epoch  16 Batch   24/26   train_loss = 0.427
Epoch  17 Batch    8/26   train_loss = 0.434
Epoch  17 Batch   18/26   train_loss = 0.434
Epoch  18 Batch    2/26   train_loss = 0.425
Epoch  18 Batch   12/26   train_loss = 0.394
Epoch  18 Batch   22/26   train_loss = 0.393
Epoch  19 Batch    6/26   train_loss = 0.367
Epoch  19 Batch   16/26   train_loss = 0.365
Epoch  20 Batch    0/26   train_loss = 0.333
Epoch  20 Batch   10/26   train_loss = 0.312
Epoch  20 Batch   20/26   train_loss = 0.306
Epoch  21 Batch    4/26   train_loss = 0.313
Epoch  21 Batch   14/26   train_loss = 0.300
Epoch  21 Batch   24/26   train_loss = 0.269
Epoch  22 Batch    8/26   train_loss = 0.275
Epoch  22 Batch   18/26   train_loss = 0.267
Epoch  23 Batch    2/26   train_loss = 0.270
Epoch  23 Batch   12/26   train_loss = 0.264
Epoch  23 Batch   22/26   train_loss = 0.259
Epoch  24 Batch    6/26   train_loss = 0.260
Epoch  24 Batch   16/26   train_loss = 0.256
Model Trained and Saved

Save Parameters

Save seq_length and save_dir for generating a new TV script.


In [17]:
"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
# Save parameters for checkpoint
helper.save_params((seq_length, save_dir))

Checkpoint


In [18]:
"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
import tensorflow as tf
import numpy as np
import helper
import problem_unittests as tests

_, vocab_to_int, int_to_vocab, token_dict = helper.load_preprocess()
seq_length, load_dir = helper.load_params()

Implement Generate Functions

Get Tensors

Get tensors from loaded_graph using the function get_tensor_by_name(). Get the tensors using the following names:

  • "input:0"
  • "initial_state:0"
  • "final_state:0"
  • "probs:0"

Return the tensors in the following tuple (InputTensor, InitialStateTensor, FinalStateTensor, ProbsTensor)


In [19]:
def get_tensors(loaded_graph):
    """
    Get input, initial state, final state, and probabilities tensor from <loaded_graph>
    :param loaded_graph: TensorFlow graph loaded from file
    :return: Tuple (InputTensor, InitialStateTensor, FinalStateTensor, ProbsTensor)
    """
    # TODO: Implement Function
    
    input_tensor =  loaded_graph.get_tensor_by_name("input:0")
    initial_state_tensor = loaded_graph.get_tensor_by_name("initial_state:0")
    final_state_tensor = loaded_graph.get_tensor_by_name("final_state:0")
    probs_tensor = loaded_graph.get_tensor_by_name("probs:0")
    
    return input_tensor, initial_state_tensor, final_state_tensor, probs_tensor


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_get_tensors(get_tensors)


Tests Passed

Choose Word

Implement the pick_word() function to select the next word using probabilities.


In [20]:
def pick_word(probabilities, int_to_vocab):
    """
    Pick the next word in the generated text
    :param probabilities: Probabilites of the next word
    :param int_to_vocab: Dictionary of word ids as the keys and words as the values
    :return: String of the predicted word
    """
    # TODO: Implement Function
    
    # idx = np.argmax(probabilities)#   greedy strategy
    idx = np.random.choice(len(int_to_vocab),p=probabilities) # reviewer suggestion
    
    return int_to_vocab[idx]


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_pick_word(pick_word)


Tests Passed

Generate TV Script

This will generate the TV script for you. Set gen_length to the length of TV script you want to generate.


In [21]:
gen_length = 200
# homer_simpson, moe_szyslak, or Barney_Gumble
prime_word = 'moe_szyslak'

"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
loaded_graph = tf.Graph()
with tf.Session(graph=loaded_graph) as sess:
    # Load saved model
    loader = tf.train.import_meta_graph(load_dir + '.meta')
    loader.restore(sess, load_dir)

    # Get Tensors from loaded model
    input_text, initial_state, final_state, probs = get_tensors(loaded_graph)

    # Sentences generation setup
    gen_sentences = [prime_word + ':']
    prev_state = sess.run(initial_state, {input_text: np.array([[1]])})

    # Generate sentences
    for n in range(gen_length):
        # Dynamic Input
        dyn_input = [[vocab_to_int[word] for word in gen_sentences[-seq_length:]]]
        dyn_seq_length = len(dyn_input[0])

        # Get Prediction
        probabilities, prev_state = sess.run(
            [probs, final_state],
            {input_text: dyn_input, initial_state: prev_state})
        
        pred_word = pick_word(probabilities[dyn_seq_length-1], int_to_vocab)

        gen_sentences.append(pred_word)
    
    # Remove tokens
    tv_script = ' '.join(gen_sentences)
    for key, token in token_dict.items():
        ending = ' ' if key in ['\n', '(', '"'] else ''
        tv_script = tv_script.replace(' ' + token.lower(), key)
    tv_script = tv_script.replace('\n ', '\n')
    tv_script = tv_script.replace('( ', '(')
        
    print(tv_script)


moe_szyslak:... so like you're looking at a. m.(nods noise, then) there's gotta be a lot happier...
lisa_simpson: you have thirteen pickled eggs in this jar!
moe_szyslak: good times...(biggest ever) whaaaa?!
moe_szyslak: my craphole! my car!
carl_carlson: all right, we're back in business.
football_announcer: so, with my daughter hates me because of us have to look at me. i'm not turn back.
moe_szyslak: really not cheerin' me up here, but uh, let me offer you the only literature in the place there. lenny, i've always wondered what it would feel like to wear something that's been ironed.
chief_wiggum: hm, only one way to be sure.
chief_wiggum:(angry) that was my insulin!
detective_homer_simpson:(haplessly) uh-oh, spaghetti-o's!
seymour_skinner:(boozy) there's something i've wanted to say to you for you.
seymour_skinner: i understand. it can be very difficult for a woman to make him a gas.

The TV Script is Nonsensical

It's ok if the TV script doesn't make any sense. We trained on less than a megabyte of text. In order to get good results, you'll have to use a smaller vocabulary or get more data. Luckly there's more data! As we mentioned in the begging of this project, this is a subset of another dataset. We didn't have you train on all the data, because that would take too long. However, you are free to train your neural network on all the data. After you complete the project, of course.

Submitting This Project

When submitting this project, make sure to run all the cells before saving the notebook. Save the notebook file as "dlnd_tv_script_generation.ipynb" and save it as a HTML file under "File" -> "Download as". Include the "helper.py" and "problem_unittests.py" files in your submission.