为什么 CNN-LSTM 比 LSTM 快？

Question

I was confused about the reason for speed up. The gain in both training and prediction speed is huge, more than 50 times.

This is how I create my LSTM model:

def create_model(learning_rate, num_LSTM_layers,
                 num_LSTM_nodes, dropout_rate):


    #CREATE THE LSTM NEURAL NETWORK
    model = Sequential()
    if num_LSTM_layers > 1:
        model.add(LSTM(num_LSTM_nodes, return_sequences=True ))
    if num_LSTM_layers == 1:
        model.add(LSTM(num_LSTM_nodes, return_sequences=False))
    model.add(Activation('relu'))
    model.add(Dropout(dropout_rate))

    if num_LSTM_layers > 1:
        for i in range(num_LSTM_layers-1):
            if i+1 == num_LSTM_layers-1:
                model.add(layers.LSTM(num_LSTM_nodes, return_sequences=False))
            else:
                model.add(layers.LSTM(num_LSTM_nodes, return_sequences=True))
            model.add(Activation('relu'))
            model.add(Dropout(dropout_rate))

    model.add(Dense(1))
    model.add(Activation('linear'))


    # Use the Adam method for training the network.
    # We want to find the best learning-rate for the Adam method.
    optimizer = Adam(lr=learning_rate)

    # In Keras we need to compile the model so it can be trained.
    model.compile(loss='mean_squared_error', optimizer=optimizer)

    return model

and this is how I create my CNN-LSTM model:

def create_model_TD(learning_rate, num_conv_layers, num_LSTM_layers,
                 num_LSTM_nodes, dropout_rate, filter_size, kernel_height, pool_size):

    #CREATE THE LSTM NEURAL NETWORK
    model = Sequential()
    model.add(TimeDistributed(Conv1D(input_shape=(None, X_train.shape[2], X_train.shape[3]) , 
                                     filters= int(filter_size), kernel_size= int(kernel_height), activation='relu', padding='causal')))
    if num_conv_layers == 2:
        model.add(TimeDistributed(Conv1D(filters=int(filter_size), kernel_size= int(kernel_height), activation='relu', padding='causal')))
    model.add(TimeDistributed(MaxPooling1D(pool_size=int(pool_size))))
    model.add(TimeDistributed(Flatten()))
    if num_LSTM_layers > 1:
        model.add(LSTM(num_LSTM_nodes, return_sequences=True))
    if num_LSTM_layers == 1:
        model.add(LSTM(num_LSTM_nodes, return_sequences=False))
    model.add(Activation('relu'))
    model.add(Dropout(dropout_rate))

    if num_LSTM_layers > 1:
        for i in range(num_LSTM_layers-1):
            if i+1 == num_LSTM_layers-1:
                model.add(LSTM(num_LSTM_nodes, return_sequences=False))
            else:
                model.add(LSTM(num_LSTM_nodes, return_sequences=True))
            model.add(Activation('relu'))
            model.add(Dropout(dropout_rate))

    model.add(Dense(1))
    model.add(Activation('linear'))


    # Use the Adam method for training the network.
    # We want to find the best learning-rate for the Adam method.
    optimizer = Adam(lr=learning_rate)

    # In Keras we need to compile the model so it can be trained.
    model.compile(loss='mean_squared_error', optimizer=optimizer)

    return model

But when I look at the number of trainable parameters, CNN-LSTM seems to have even more parameters than classic LSTM. Anyone knows the reason? I would appreciate your help, thanks.

Answer 1

CNN leads to a variety of different complexity reductions by concentrating on the key features. The use of convolution layers leads to a reduction in the size of tensors. Also the use of pooling leads to a further reduction. And last but not least the ReLu layer reduces the complexity. Because of the training time decrease

Deeper beginner look in cnn

https://www.semanticscholar.org/paper/Introduction-to-Convolutional-Neural-Networks-Wu/450ca19932fcef1ca6d0442cbf52fec38fb9d1e5