word_index = {k:v for k,v in tokenizer.word_index.items()}
word_index["<PAD>"] = 0

vocab_size = len(word_index)

maxLen = length_longest_sentence
data = tf.keras.preprocessing.sequence.pad_sequences(sent_numeric,
                                                        value=word_index["<PAD>"],
                                                        padding='post',
                                                        maxlen=maxLen)