ValueError：层“conv_lstm1d”的输入 0 与该层不兼容：预期 ndim=4，发现 ndim=3。收到完整形状：（无、300、17）

Question

我需要帮助来找出这段代码中缺少的内容。

这是模型：

def make_model(input_shape):
    input_layer = keras.layers.Input(input_shape)

    convLSTM1 = keras.layers.ConvLSTM1D(filters=32, kernel_size=sfreq, strides=2, padding="valid")(input_layer)
    convLSTM1 = keras.layers.BatchNormalization()(convLSTM1)
    convLSTM1 = keras.layers.ReLU()(convLSTM1)

    convLSTM2 = keras.layers.ConvLSTM1D(filters=64, kernel_size=sfreq, strides=2,padding="valid")(conv1)
    convLSTM2 = keras.layers.BatchNormalization()(convLSTM2)
    convLSTM2 = keras.layers.ReLU()(convLSTM2)

    convLSTM3 = keras.layers.ConvLSTM1D(filters=128, kernel_size=sfreq, strides=2,padding="valid")(conv2)
    convLSTM3 = keras.layers.BatchNormalization()(convLSTM3)
    convLSTM3 = keras.layers.ReLU()(convLSTM3)

    gap = keras.layers.GlobalAveragePooling1D()(convLSTM3)

    output_layer = keras.layers.Dense(num_classes, activation="softmax")(gap)

    return keras.models.Model(inputs=input_layer, outputs=output_layer)


model = make_model(input_shape=X_train.shape[1:])
model.summary()

出现此错误：

ValueError: Input 0 of layer "conv_lstm1d" is incompatible with the layer: expected ndim=4, found ndim=3. Full shape received: (None, 300, 17)

Answer 1

您的代码中有一个类型

convLSTM2 = keras.layers.ConvLSTM1D(filters=64, kernel_size=sfreq, strides=2,padding="valid")(conv1)

应该是

convLSTM2 = keras.layers.ConvLSTM1D(filters=64, kernel_size=sfreq, strides=2,padding="valid")(conv1LSTM1)

第 3 层代码中也有类似的拼写错误