tensorflow 2.0 循環神經網絡 之 情感分類實戰

兩種方法 （IMDB數據集）

• SimpleRNNCell （易理解其原理）

  • single layer
  • multi-layers

• RNNCell

IMDB 數據集加載

1. 加載 IMDB 數據集。

# the most frequest words
total_words = 10000  # 常用單詞數 10000 其他用統一標記符來表示。
max_review_len = 80   # 句子長度 80
embedding_len = 100
(x_train, y_train), (x_test, y_test) = keras.datasets.imdb.load_data(num_words=total_words)
# x_train:[b, 80]
# x_test: [b, 80]
x_train = keras.preprocessing.sequence.pad_sequences(x_train, maxlen=max_review_len)
x_test = keras.preprocessing.sequence.pad_sequences(x_test, maxlen=max_review_len)

db_train = tf.data.Dataset.from_tensor_slices((x_train, y_train))
db_train = db_train.shuffle(1000).batch(batchsz, drop_remainder=True)
db_test = tf.data.Dataset.from_tensor_slices((x_test, y_test))
db_test = db_test.batch(batchsz, drop_remainder=True)
print('x_train shape:', x_train.shape, tf.reduce_max(y_train), tf.reduce_min(y_train))
print('x_test shape:', x_test.shape)

x_train shape: (25000, 80) tf.Tensor(1, shape=(), dtype=int64) tf.Tensor(0, shape=(), dtype=int64)
x_test shape: (25000, 80)

定義 RNN

1. 兩個 cell 循環 形成 layer。

class MyRNN(keras.Model):

    def __init__(self, units):
        super(MyRNN, self).__init__()

        # [b, 64]
        self.state0 = [tf.zeros([batchsz, units])]
        self.state1 = [tf.zeros([batchsz, units])]

        # transform text to embedding representation
        # [b, 80] => [b, 80, 100]
        self.embedding = layers.Embedding(total_words, embedding_len,
                                          input_length=max_review_len)

        # [b, 80, 100] , h_dim: 64
        # RNN: cell1 ,cell2, cell3
        # SimpleRNN
        self.rnn_cell0 = layers.SimpleRNNCell(units, dropout=0.5)
        self.rnn_cell1 = layers.SimpleRNNCell(units, dropout=0.5)


        # fc, [b, 80, 100] => [b, 64] => [b, 1]
        self.outlayer = layers.Dense(1)

    def call(self, inputs, training=None):
        """
        net(x) net(x, training=True) :train mode
        net(x, training=False): test
        :param inputs: [b, 80]
        :param training:
        :return:
        """
        # [b, 80]
        x = inputs
        # embedding: [b, 80] => [b, 80, 100]
        x = self.embedding(x)
        # rnn cell compute
        # [b, 80, 100] => [b, 64]
        state0 = self.state0
        state1 = self.state1
        for word in tf.unstack(x, axis=1): # word: [b, 100]
            # h1 = x*wxh+h0*whh
            # out0: [b, 64]
            out0, state0 = self.rnn_cell0(word, state0, training)
            # out1: [b, 64]
            out1, state1 = self.rnn_cell1(out0, state1, training)

        # out: [b, 64] => [b, 1]
        x = self.outlayer(out1)
        # p(y is pos|x)
        prob = tf.sigmoid(x)

        return prob

訓練測試

def main():
    units = 64
    epochs = 3

    model = MyRNN(units)
    model.compile(optimizer = keras.optimizers.Adam(0.001),
                  loss = tf.losses.BinaryCrossentropy(),
                  metrics=['accuracy'])
    model.fit(db_train, epochs=epochs, validation_data=db_test)

    print('test:')
    model.evaluate(db_test)

結果：

train：
loss: 0.3122 - accuracy: 0.8709
test：
loss: 0.4019 - accuracy: 0.8301