Classificação de texto com um RNN

Ver no TensorFlow.org Executar no Google Colab Ver fonte no GitHub Baixar caderno

Este tutorial classificação texto treina uma rede neural recorrente no IMDB grande filme conjunto de dados revisão para análise de sentimento.

Configurar

import numpy as np

import tensorflow_datasets as tfds
import tensorflow as tf

tfds.disable_progress_bar()

Import matplotlib e criar uma função auxiliar para gráficos enredo:

import matplotlib.pyplot as plt


def plot_graphs(history, metric):
  plt.plot(history.history[metric])
  plt.plot(history.history['val_'+metric], '')
  plt.xlabel("Epochs")
  plt.ylabel(metric)
  plt.legend([metric, 'val_'+metric])

Configurar pipeline de entrada

O IMDB grande revisão de filme conjunto de dados é um binário de classificação do conjunto de dados-todos os comentários têm uma positiva ou sentimento negativo.

Faça o download do conjunto de dados usando TFDS . Veja o tutorial texto de carregamento para obter detalhes sobre como carregar este tipo de dados manualmente.

dataset, info = tfds.load('imdb_reviews', with_info=True,
                          as_supervised=True)
train_dataset, test_dataset = dataset['train'], dataset['test']

train_dataset.element_spec
(TensorSpec(shape=(), dtype=tf.string, name=None),
 TensorSpec(shape=(), dtype=tf.int64, name=None))

Inicialmente, isso retorna um conjunto de dados de (texto, pares de rótulos):

for example, label in train_dataset.take(1):
  print('text: ', example.numpy())
  print('label: ', label.numpy())
text:  b"This was an absolutely terrible movie. Don't be lured in by Christopher Walken or Michael Ironside. Both are great actors, but this must simply be their worst role in history. Even their great acting could not redeem this movie's ridiculous storyline. This movie is an early nineties US propaganda piece. The most pathetic scenes were those when the Columbian rebels were making their cases for revolutions. Maria Conchita Alonso appeared phony, and her pseudo-love affair with Walken was nothing but a pathetic emotional plug in a movie that was devoid of any real meaning. I am disappointed that there are movies like this, ruining actor's like Christopher Walken's good name. I could barely sit through it."
label:  0

Próxima embaralhar os dados de formação e criar lotes destes (text, label) pares:

BUFFER_SIZE = 10000
BATCH_SIZE = 64
train_dataset = train_dataset.shuffle(BUFFER_SIZE).batch(BATCH_SIZE).prefetch(tf.data.AUTOTUNE)
test_dataset = test_dataset.batch(BATCH_SIZE).prefetch(tf.data.AUTOTUNE)
for example, label in train_dataset.take(1):
  print('texts: ', example.numpy()[:3])
  print()
  print('labels: ', label.numpy()[:3])
texts:  [b'This is arguably the worst film I have ever seen, and I have quite an appetite for awful (and good) movies. It could (just) have managed a kind of adolescent humour if it had been consistently tongue-in-cheek --\xc3\xa0 la ROCKY HORROR PICTURE SHOW, which was really very funny. Other movies, like PLAN NINE FROM OUTER SPACE, manage to be funny while (apparently) trying to be serious. As to the acting, it looks like they rounded up brain-dead teenagers and asked them to ad-lib the whole production. Compared to them, Tom Cruise looks like Alec Guinness. There was one decent interpretation -- that of the older ghoul-busting broad on the motorcycle.'
 b"I saw this film in the worst possible circumstance. I'd already missed 15 minutes when I woke up to it on an international flight between Sydney and Seoul. I didn't know what I was watching, I thought maybe it was a movie of the week, but quickly became riveted by the performance of the lead actress playing a young woman who's child had been kidnapped. The premise started taking twist and turns I didn't see coming and by the end credits I was scrambling through the the in-flight guide to figure out what I had just watched. Turns out I was belatedly discovering Do-yeon Jeon who'd won Best Actress at Cannes for the role. I don't know if Secret Sunshine is typical of Korean cinema but I'm off to the DVD store to discover more."
 b"Hello. I am Paul Raddick, a.k.a. Panic Attack of WTAF, Channel 29 in Philadelphia. Let me tell you about this god awful movie that powered on Adam Sandler's film career but was digitized after a short time.<br /><br />Going Overboard is about an aspiring comedian played by Sandler who gets a job on a cruise ship and fails...or so I thought. Sandler encounters babes that like History of the World Part 1 and Rebound. The babes were supposed to be engaged, but, actually, they get executed by Sawtooth, the meanest cannibal the world has ever known. Adam Sandler fared bad in Going Overboard, but fared better in Big Daddy, Billy Madison, and Jen Leone's favorite, 50 First Dates. Man, Drew Barrymore was one hot chick. Spanglish is red hot, Going Overboard ain't Dooley squat! End of file."]

labels:  [0 1 0]

Crie o codificador de texto

O texto cru carregado pelo tfds precisa ser processado antes que ele possa ser usado em um modelo. A maneira mais simples de texto processo de formação está usando o TextVectorization camada. Esta camada tem muitos recursos, mas este tutorial mantém o comportamento padrão.

Criar a camada, e passar o texto do conjunto de dados para a camada .adapt método:

VOCAB_SIZE = 1000
encoder = tf.keras.layers.TextVectorization(
    max_tokens=VOCAB_SIZE)
encoder.adapt(train_dataset.map(lambda text, label: text))

O .adapt método define o vocabulário da camada. Aqui estão os primeiros 20 tokens. Após o preenchimento e os tokens desconhecidos, eles são classificados por frequência:

vocab = np.array(encoder.get_vocabulary())
vocab[:20]
array(['', '[UNK]', 'the', 'and', 'a', 'of', 'to', 'is', 'in', 'it', 'i',
       'this', 'that', 'br', 'was', 'as', 'for', 'with', 'movie', 'but'],
      dtype='<U14')

Depois que o vocabulário é definido, a camada pode codificar o texto em índices. Os tensores de índices são 0-impregnado com a sequência mais longa do lote (a menos que defina uma fixo output_sequence_length ):

encoded_example = encoder(example)[:3].numpy()
encoded_example
array([[ 11,   7,   1, ...,   0,   0,   0],
       [ 10, 208,  11, ...,   0,   0,   0],
       [  1,  10, 237, ...,   0,   0,   0]])

Com as configurações padrão, o processo não é totalmente reversível. Existem três razões principais para isso:

  1. O valor padrão para preprocessing.TextVectorization 's standardize argumento é "lower_and_strip_punctuation" .
  2. O tamanho limitado do vocabulário e a falta de fallback baseado em caracteres resultam em alguns tokens desconhecidos.
for n in range(3):
  print("Original: ", example[n].numpy())
  print("Round-trip: ", " ".join(vocab[encoded_example[n]]))
  print()
Original:  b'This is arguably the worst film I have ever seen, and I have quite an appetite for awful (and good) movies. It could (just) have managed a kind of adolescent humour if it had been consistently tongue-in-cheek --\xc3\xa0 la ROCKY HORROR PICTURE SHOW, which was really very funny. Other movies, like PLAN NINE FROM OUTER SPACE, manage to be funny while (apparently) trying to be serious. As to the acting, it looks like they rounded up brain-dead teenagers and asked them to ad-lib the whole production. Compared to them, Tom Cruise looks like Alec Guinness. There was one decent interpretation -- that of the older ghoul-busting broad on the motorcycle.'
Round-trip:  this is [UNK] the worst film i have ever seen and i have quite an [UNK] for awful and good movies it could just have [UNK] a kind of [UNK] [UNK] if it had been [UNK] [UNK] [UNK] la [UNK] horror picture show which was really very funny other movies like [UNK] [UNK] from [UNK] space [UNK] to be funny while apparently trying to be serious as to the acting it looks like they [UNK] up [UNK] [UNK] and [UNK] them to [UNK] the whole production [UNK] to them tom [UNK] looks like [UNK] [UNK] there was one decent [UNK] that of the older [UNK] [UNK] on the [UNK]                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               

Original:  b"I saw this film in the worst possible circumstance. I'd already missed 15 minutes when I woke up to it on an international flight between Sydney and Seoul. I didn't know what I was watching, I thought maybe it was a movie of the week, but quickly became riveted by the performance of the lead actress playing a young woman who's child had been kidnapped. The premise started taking twist and turns I didn't see coming and by the end credits I was scrambling through the the in-flight guide to figure out what I had just watched. Turns out I was belatedly discovering Do-yeon Jeon who'd won Best Actress at Cannes for the role. I don't know if Secret Sunshine is typical of Korean cinema but I'm off to the DVD store to discover more."
Round-trip:  i saw this film in the worst possible [UNK] id already [UNK] [UNK] minutes when i [UNK] up to it on an [UNK] [UNK] between [UNK] and [UNK] i didnt know what i was watching i thought maybe it was a movie of the [UNK] but quickly became [UNK] by the performance of the lead actress playing a young woman whos child had been [UNK] the premise started taking twist and turns i didnt see coming and by the end credits i was [UNK] through the the [UNK] [UNK] to figure out what i had just watched turns out i was [UNK] [UNK] [UNK] [UNK] [UNK] [UNK] best actress at [UNK] for the role i dont know if secret [UNK] is typical of [UNK] cinema but im off to the dvd [UNK] to [UNK] more                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     

Original:  b"Hello. I am Paul Raddick, a.k.a. Panic Attack of WTAF, Channel 29 in Philadelphia. Let me tell you about this god awful movie that powered on Adam Sandler's film career but was digitized after a short time.<br /><br />Going Overboard is about an aspiring comedian played by Sandler who gets a job on a cruise ship and fails...or so I thought. Sandler encounters babes that like History of the World Part 1 and Rebound. The babes were supposed to be engaged, but, actually, they get executed by Sawtooth, the meanest cannibal the world has ever known. Adam Sandler fared bad in Going Overboard, but fared better in Big Daddy, Billy Madison, and Jen Leone's favorite, 50 First Dates. Man, Drew Barrymore was one hot chick. Spanglish is red hot, Going Overboard ain't Dooley squat! End of file."
Round-trip:  [UNK] i am paul [UNK] [UNK] [UNK] [UNK] of [UNK] [UNK] [UNK] in [UNK] let me tell you about this god awful movie that [UNK] on [UNK] [UNK] film career but was [UNK] after a short [UNK] br going [UNK] is about an [UNK] [UNK] played by [UNK] who gets a job on a [UNK] [UNK] and [UNK] so i thought [UNK] [UNK] [UNK] that like history of the world part 1 and [UNK] the [UNK] were supposed to be [UNK] but actually they get [UNK] by [UNK] the [UNK] [UNK] the world has ever known [UNK] [UNK] [UNK] bad in going [UNK] but [UNK] better in big [UNK] [UNK] [UNK] and [UNK] [UNK] favorite [UNK] first [UNK] man [UNK] [UNK] was one hot [UNK] [UNK] is red hot going [UNK] [UNK] [UNK] [UNK] end of [UNK]

Crie o modelo

Um desenho do fluxo de informações no modelo

Acima está um diagrama do modelo.

  1. Este modelo pode ser construído como um tf.keras.Sequential .

  2. A primeira camada é o encoder , o qual converte o texto para uma sequência de índices de token.

  3. Depois que o codificador é uma camada de incorporação. Uma camada de incorporação armazena um vetor por palavra. Quando chamado, ele converte as sequências de índices de palavras em sequências de vetores. Esses vetores são treináveis. Após o treinamento (com dados suficientes), palavras com significados semelhantes geralmente têm vetores semelhantes.

    Este índice de pesquisa é muito mais eficiente do que a operação equivalente de passar um vector codificado um-a quente através de um tf.keras.layers.Dense camada.

  4. Uma rede neural recorrente (RNN) processa a entrada de sequência iterando através dos elementos. Os RNNs passam as saídas de um passo de tempo para sua entrada no próximo passo de tempo.

    O tf.keras.layers.Bidirectional invólucro também pode ser usado com uma camada de RNN. Isso propaga a entrada para a frente e para trás através da camada RNN e então concatena a saída final.

    • A principal vantagem de um RNN bidirecional é que o sinal do início da entrada não precisa ser processado em todos os passos de tempo para afetar a saída.

    • A principal desvantagem de um RNN bidirecional é que você não pode transmitir previsões de maneira eficiente à medida que as palavras são adicionadas ao final.

  5. Após a RNN converteu a sequência de um único vector dos dois layers.Dense fazer algum processamento final, e convertido a partir desta representação do vector para um único logit como a saída de classificação.

O código para implementar isso está abaixo:

model = tf.keras.Sequential([
    encoder,
    tf.keras.layers.Embedding(
        input_dim=len(encoder.get_vocabulary()),
        output_dim=64,
        # Use masking to handle the variable sequence lengths
        mask_zero=True),
    tf.keras.layers.Bidirectional(tf.keras.layers.LSTM(64)),
    tf.keras.layers.Dense(64, activation='relu'),
    tf.keras.layers.Dense(1)
])

Observe que o modelo sequencial Keras é usado aqui, uma vez que todas as camadas no modelo têm apenas uma entrada e produzem uma saída única. Caso queira usar a camada RNN com estado, você pode querer construir seu modelo com a API funcional Keras ou subclasse de modelo para que possa recuperar e reutilizar os estados da camada RNN. Verifique guia Keras RNN para mais detalhes.

As camadas de embebimento utiliza o mascaramento para lidar com as variáveis de sequências de comprimentos. Todas as camadas após a Embedding mascaramento apoio:

print([layer.supports_masking for layer in model.layers])
[False, True, True, True, True]

Para confirmar se isso funciona conforme o esperado, avalie uma frase duas vezes. Primeiro, sozinho, para que não haja preenchimento para mascarar:

# predict on a sample text without padding.

sample_text = ('The movie was cool. The animation and the graphics '
               'were out of this world. I would recommend this movie.')
predictions = model.predict(np.array([sample_text]))
print(predictions[0])
[-0.00012211]

Agora, avalie novamente em um lote com uma frase mais longa. O resultado deve ser idêntico:

# predict on a sample text with padding

padding = "the " * 2000
predictions = model.predict(np.array([sample_text, padding]))
print(predictions[0])
[-0.00012211]

Compile o modelo Keras para configurar o processo de treinamento:

model.compile(loss=tf.keras.losses.BinaryCrossentropy(from_logits=True),
              optimizer=tf.keras.optimizers.Adam(1e-4),
              metrics=['accuracy'])

Treine o modelo

history = model.fit(train_dataset, epochs=10,
                    validation_data=test_dataset,
                    validation_steps=30)
Epoch 1/10
391/391 [==============================] - 39s 84ms/step - loss: 0.6454 - accuracy: 0.5630 - val_loss: 0.4888 - val_accuracy: 0.7568
Epoch 2/10
391/391 [==============================] - 30s 75ms/step - loss: 0.3925 - accuracy: 0.8200 - val_loss: 0.3663 - val_accuracy: 0.8464
Epoch 3/10
391/391 [==============================] - 30s 75ms/step - loss: 0.3319 - accuracy: 0.8525 - val_loss: 0.3402 - val_accuracy: 0.8385
Epoch 4/10
391/391 [==============================] - 30s 75ms/step - loss: 0.3183 - accuracy: 0.8616 - val_loss: 0.3289 - val_accuracy: 0.8438
Epoch 5/10
391/391 [==============================] - 30s 75ms/step - loss: 0.3088 - accuracy: 0.8656 - val_loss: 0.3254 - val_accuracy: 0.8646
Epoch 6/10
391/391 [==============================] - 32s 81ms/step - loss: 0.3043 - accuracy: 0.8686 - val_loss: 0.3242 - val_accuracy: 0.8521
Epoch 7/10
391/391 [==============================] - 30s 76ms/step - loss: 0.3019 - accuracy: 0.8696 - val_loss: 0.3315 - val_accuracy: 0.8609
Epoch 8/10
391/391 [==============================] - 32s 76ms/step - loss: 0.3007 - accuracy: 0.8688 - val_loss: 0.3245 - val_accuracy: 0.8609
Epoch 9/10
391/391 [==============================] - 31s 77ms/step - loss: 0.2981 - accuracy: 0.8707 - val_loss: 0.3294 - val_accuracy: 0.8599
Epoch 10/10
391/391 [==============================] - 31s 78ms/step - loss: 0.2969 - accuracy: 0.8742 - val_loss: 0.3218 - val_accuracy: 0.8547
test_loss, test_acc = model.evaluate(test_dataset)

print('Test Loss:', test_loss)
print('Test Accuracy:', test_acc)
391/391 [==============================] - 15s 38ms/step - loss: 0.3185 - accuracy: 0.8582
Test Loss: 0.3184521794319153
Test Accuracy: 0.8581600189208984
plt.figure(figsize=(16, 8))
plt.subplot(1, 2, 1)
plot_graphs(history, 'accuracy')
plt.ylim(None, 1)
plt.subplot(1, 2, 2)
plot_graphs(history, 'loss')
plt.ylim(0, None)
(0.0, 0.6627909764647484)

png

Faça uma previsão em uma nova frase:

Se a previsão for> = 0,0, é positiva, caso contrário, é negativa.

sample_text = ('The movie was cool. The animation and the graphics '
               'were out of this world. I would recommend this movie.')
predictions = model.predict(np.array([sample_text]))

Empilhe duas ou mais camadas LSTM

Camadas recorrentes Keras tem dois modos disponíveis que são controlados pela return_sequences argumento do construtor:

  • Se False ele retorna somente o último de saída para cada sequência de entrada (um tensor de forma 2D (batch_size, output_features)). Este é o padrão, usado no modelo anterior.

  • Se True as sequências completas de saídas sucessivas para cada iteração temporal é devolvido (um tensor 3D de forma (batch_size, timesteps, output_features) ).

Aqui está o que o fluxo de olhares de informação como com return_sequences=True :

layered_bidirectional

A coisa interessante sobre o uso de uma RNN com return_sequences=True é que a saída ainda tem 3 eixos, como a entrada, para que possa ser passado para outra camada RNN, como este:

model = tf.keras.Sequential([
    encoder,
    tf.keras.layers.Embedding(len(encoder.get_vocabulary()), 64, mask_zero=True),
    tf.keras.layers.Bidirectional(tf.keras.layers.LSTM(64,  return_sequences=True)),
    tf.keras.layers.Bidirectional(tf.keras.layers.LSTM(32)),
    tf.keras.layers.Dense(64, activation='relu'),
    tf.keras.layers.Dropout(0.5),
    tf.keras.layers.Dense(1)
])
model.compile(loss=tf.keras.losses.BinaryCrossentropy(from_logits=True),
              optimizer=tf.keras.optimizers.Adam(1e-4),
              metrics=['accuracy'])
history = model.fit(train_dataset, epochs=10,
                    validation_data=test_dataset,
                    validation_steps=30)
Epoch 1/10
391/391 [==============================] - 71s 149ms/step - loss: 0.6502 - accuracy: 0.5625 - val_loss: 0.4923 - val_accuracy: 0.7573
Epoch 2/10
391/391 [==============================] - 55s 138ms/step - loss: 0.4067 - accuracy: 0.8198 - val_loss: 0.3727 - val_accuracy: 0.8271
Epoch 3/10
391/391 [==============================] - 54s 136ms/step - loss: 0.3417 - accuracy: 0.8543 - val_loss: 0.3343 - val_accuracy: 0.8510
Epoch 4/10
391/391 [==============================] - 53s 134ms/step - loss: 0.3242 - accuracy: 0.8607 - val_loss: 0.3268 - val_accuracy: 0.8568
Epoch 5/10
391/391 [==============================] - 53s 135ms/step - loss: 0.3174 - accuracy: 0.8652 - val_loss: 0.3213 - val_accuracy: 0.8516
Epoch 6/10
391/391 [==============================] - 52s 132ms/step - loss: 0.3098 - accuracy: 0.8671 - val_loss: 0.3294 - val_accuracy: 0.8547
Epoch 7/10
391/391 [==============================] - 53s 134ms/step - loss: 0.3063 - accuracy: 0.8697 - val_loss: 0.3158 - val_accuracy: 0.8594
Epoch 8/10
391/391 [==============================] - 52s 132ms/step - loss: 0.3043 - accuracy: 0.8692 - val_loss: 0.3184 - val_accuracy: 0.8521
Epoch 9/10
391/391 [==============================] - 53s 133ms/step - loss: 0.3016 - accuracy: 0.8704 - val_loss: 0.3208 - val_accuracy: 0.8609
Epoch 10/10
391/391 [==============================] - 54s 136ms/step - loss: 0.2975 - accuracy: 0.8740 - val_loss: 0.3301 - val_accuracy: 0.8651
test_loss, test_acc = model.evaluate(test_dataset)

print('Test Loss:', test_loss)
print('Test Accuracy:', test_acc)
391/391 [==============================] - 26s 65ms/step - loss: 0.3293 - accuracy: 0.8646
Test Loss: 0.329334557056427
Test Accuracy: 0.8646399974822998
# predict on a sample text without padding.

sample_text = ('The movie was not good. The animation and the graphics '
               'were terrible. I would not recommend this movie.')
predictions = model.predict(np.array([sample_text]))
print(predictions)
[[-1.6796288]]
plt.figure(figsize=(16, 6))
plt.subplot(1, 2, 1)
plot_graphs(history, 'accuracy')
plt.subplot(1, 2, 2)
plot_graphs(history, 'loss')

png

Confira outras camadas recorrentes existentes, tais como camadas GRU .

Se você está interestied na construção RNNs personalizados, consulte o Guia de RNN Keras .