输入 0 与层 lstm_16 不兼容：预期 ndim=3，发现 ndim=2？

Question

我收到错误Input 0 is incompatible with layer lstm_16: expected ndim=3, found ndim=2，代码如下：

#Step 6: Initialize the RNN
regressor = Sequential()

#Step 7: Adding the LSTM layers and some Dropout regularization

#Dropout regularization: drops out unnecessary data, so we are not shifting huge amounts of data through the network

#Adding the first LSTM layer and some Dropout regularization
#units - dimensionality in output space
regressor.add(LSTM(units = 50, return_sequences = True, input_shape = (X_train.shape[1], 1)))

#Adding the second LSTM layer and some Dropout regulariation
regressor.add(LSTM(units = 25, return_sequences = True))
regressor.add(Dropout(0.2))

#Adding a third LSTM layer and some Dropout regularization
regressor.add(LSTM(units = 12, return_sequences = True))
regressor.add(Dropout(0.2))

#Adding a fourth LSTM layer and some Dropout regularization
regressor.add(LSTM(units = 6))
regressor.add(Dropout(0.2))

#Step 8: Adding the output layer
regressor.add(Dense(units = 1))

X_train 的尺寸为(114, 1, 216)（我电脑时产生的X_train.shape）。为什么会出现此错误？

Answer 1

在这里，您在第一个 LSTM 层的 input_shape 中传递两个值。

Keras LSTM 接受并输入 (samples, time_steps, nfeatures) 的形状，并且您的图层输入必须具有此形状。