How to use the pretrained model in an application? - tensorflow

Question

I dont understand how to use the trained model.
For example I trained the model using the code from https://towardsdatascience.com/what-is-a-recurrent-neural-network-and-implementation-of-simplernn-gru-and-lstm-models-in-keras-f7247e97c405.
import numpy as np
import tensorflow as tf
import tensorflow_datasets as tfds
from tensorflow.keras.preprocessing.text import Tokenizer
from tensorflow.keras.preprocessing.sequence import pad_sequences
imdb, info = tfds.load("imdb_reviews",with_info=True, as_supervised=True)
train_data, test_data = imdb['train'], imdb['test']
training_sentences = []
training_labels = []
testing_sentences = []
testing_labels = []
for s,l in train_data:
training_sentences.append(str(s.numpy()))
training_labels.append(l.numpy())
for s,l in test_data:
testing_sentences.append(str(s.numpy()))
testing_labels.append(l.numpy())
training_labels_final = np.array(training_labels)
testing_labels_final = np.array(testing_labels)
vocab_size = 10000
embedding_dim=16
max_length = 120
trunc_type= 'post'
oov_tok="<OOV>"
tokenizer = Tokenizer(num_words = vocab_size, oov_token=oov_tok)
tokenizer.fit_on_texts(training_sentences)
word_index = tokenizer.word_index
word_index
sequences = tokenizer.texts_to_sequences(training_sentences)
padded = pad_sequences(sequences, maxlen=max_length,
truncating = trunc_type)
testing_sequences = tokenizer.texts_to_sequences(testing_sentences)
testing_padded = pad_sequences(testing_sequences, maxlen=max_length)
model = tf.keras.Sequential([
tf.keras.layers.Embedding(vocab_size, embedding_dim,
input_length=max_length),
tf.keras.layers.SimpleRNN(32),
tf.keras.layers.Dense(10, activation='relu'),
tf.keras.layers.Dense(1, activation='sigmoid')
])
model.summary()
model.compile(loss='binary_crossentropy',optimizer='adam',metrics=['accuracy'])
num_epochs=30
history=model.fit(padded, training_labels_final, epochs=num_epochs, validation_data = (testing_padded, testing_labels_final))
It is all from this link.
Then I added the last line of code
model.save('name.model')
Then I open the new script where I want to use this model. I suggests that I will put a sentence ("He is a bad cat.") and I will get 0 or 1 depending wether it represents good sentiment or bad sentiment (I think I must get 0).
import tensorflow as tf
from tensorflow import keras
import numpy as np
from tensorflow.keras.preprocessing.text import Tokenizer
from tensorflow.keras.preprocessing.sequence import pad_sequences
model=tf.keras.models.load_model('name.model')
print(model.output_shape)
prediction=model.predict(["He is a bad cat."])
print(prediction)
And I get error.
QUESTION 1: In which format I must put this sentance? If I give this model to my friend how does he know in which format he must put the sentence to this model?
QUESTION 2: I noticed that the output will have (None,1) format. But I hoped to see one number (0 or 1) but not strange vector. What is going and how to get 0 or 1?

Answer 1

Before you make predictions you have to convert your text to input format.
For example, to make a prediction of "He is a bad cat." you have to tokenize the sentence.
#tokenizing the prediction data
sentence=['He is a bad cat.']
sequences=tokenizer.texts_to_sequences(sentence)
Then you have to add padding to it
#padding the tokenized sentence
padded=pad_sequences(sequences,maxlen=max_length,truncating=trunc_type)
Then you can make predictions on the padded sequence
model1.predict(padded) #output: array([[0.00079787]], dtype=float32)
The output will be a float not exactly 0 or 1 because the model will give the value based on the given sentence close to the negative or positive side on the embedding.
For more details please refer to this working gist. Thank You.