Kerastuner:'ValueError: not a legal parameter' 当我使用 LSTM 网络时出现问题，但密集层工作正常

如何解决Kerastuner:'ValueError: not a legal parameter' 当我使用 LSTM 网络时出现问题，但密集层工作正常

'ValueError: not a legal parameter' 使用 LSTM 时出现问题。但是，如果我只使用 Dense 层，它工作正常。

使用 LSTM 时发生错误。

def model_builder(hp_units1=40):
    model = Sequential()
    model.add(LSTM(units = hp_units1,return_sequences = True,input_shape = (X.shape[1],1)))
    model.add(Dropout(0.2))
    model.add(LSTM(units = 40,return_sequences = True))
    model.add(Dropout(0.2))
    model.add(Dense(units = 1))
    optimizer = Adam(learning_rate=hp_learning_rate,momentum=hp_momentum)
    model.compile(optimizer = optimizer,loss = 'mean_squared_error',metrics=['accuracy'])
    return model

hp_units1 = [30+i*5 for i in range(5)]
hp_learning_rate =[0.01,0.001,0.0001,0.00001]
hp_momentum = [0.0,0.2,0.4,0.6,0.8,0.9]
param_grid = dict(units1=hp_units1,learning_rate=hp_learning_rate,momentum=hp_momentum)

model = KerasRegressor(build_fn=model_builder,nb_epoch=100,batch_size=32,verbose=0)
grid = gridsearchcv(estimator=model,param_grid=param_grid,n_jobs=-1)
grid_result = grid.fit(X,Y)

但是，使用 Dense 层效果很好。

def baseline_model(learn_rate=0.01,momentum=0): # (optimizer = 'adam')
    model = Sequential()
    input = X.shape[1]
    model.add(Dense(5,input_shape=(input,),kernel_initializer='normal',activation='relu'))
    model.add(Dense(4,activation='linear'))
    model.add(Dense(1))
    optimizer = SGD(lr=learn_rate,momentum=momentum)
    model.compile(loss='mean_squared_error',optimizer=optimizer,metrics=['accuracy'])
    return model

model = KerasRegressor(build_fn=baseline_model,batch_size=10,verbose=0)

learn_rate = [0.001,0.01,0.1,0.3]
momentum = [0.0,0.9]
param_grid = dict(learn_rate=learn_rate,momentum=momentum)

grid = gridsearchcv(estimator=model,Y)