Forecast.model（）对于所有输入总是返回相同的结果

Question

我正在研究多元LSTM代码。当我训练模型时，

设计网络

model = Sequential()
model.add(LSTM(100, input_shape=(train_X.shape[1], train_X.shape[2])))
model.add(Dense(n_out*n_features))
model.compile(loss='mae', optimizer='adam')
# fit network
history = model.fit(train_X, train_y, epochs=10, batch_size=72, validation_data=(test_X, test_y), verbose=2, shuffle=False)


# make a prediction
yhat = model.predict(test_X)
print(yhat.shape)
print(yhat)

我注意到它只提供一组重复的值。

输出（yhat）

[[ 3.4039806e+01  1.2757405e+00  1.0543864e+01 ...  6.5443869e+00
  -4.8950994e-03  5.8005480e-03]
 [ 3.4039803e+01  1.2757403e+00  1.0543863e+01 ...  6.5443869e+00
  -4.8949206e-03  5.8003617e-03]
 [ 3.4039803e+01  1.2757401e+00  1.0543863e+01 ...  6.5443869e+00
  -4.8948163e-03  5.8003096e-03]
 ...
 [ 3.4039806e+01  1.2757384e+00  1.0543867e+01 ...  6.5443854e+00
  -4.8948908e-03  5.8027236e-03]
 [ 3.4039806e+01  1.2757379e+00  1.0543868e+01 ...  6.5443850e+00
  -4.8948014e-03  5.8030440e-03]
 [ 3.4039806e+01  1.2757375e+00  1.0543868e+01 ...  6.5443850e+00
  -4.8947120e-03  5.8032898e-03]]

如您所见，它为所有输入返回相同的输出。 我已经验证了训练和测试集，而且已经按照原样进行。