我实现了以下代码。它在之前版本的Keras中成功工作:
max_sequence = 56
input_dim = 26
print("Build model..1")
first_input = Input(shape=(max_sequence,input_dim))
first_lstm = LSTM(5, return_sequences=True)(first_input)
first_bn = BatchNormalization()(first_lstm)
first_activation = Activation('tanh')(first_bn)
first_flat = Flatten()(first_activation)
print("Build model..2")
second_input = Input(shape=(max_sequence,input_dim))
second_lstm = LSTM(5, return_sequences=True)(second_input)
second_bn = BatchNormalization()(second_lstm)
second_activation = Activation('tanh')(second_bn)
second_flat = Flatten()(second_activation)
merge=concatenate([first_flat, second_flat])
merge_dense=Dense(3)(merge)
merge_bn = BatchNormalization()(merge_dense)
merge_activation = Activation('tanh')(merge_bn)
merge_dense2=Dense(1)(merge_activation)
merge_activation2 = Activation('tanh')(merge_dense2)
train_x_1 = np.reshape(np.array(train_x_1), [2999, 56, 26])
train_x_2 = np.reshape(np.array(train_x_2), [2999, 56, 26])
model=Model(inputs=[train_x_1,train_x_2], outputs=train_y_class)
optimizer = RMSprop(lr=0.5)
model.compile(optimizer=optimizer, loss='binary_crossentropy', metrics=['accuracy'])
history = model.fit([train_x_1, train_x_2], train_y_class, nb_epoch=300, batch_size=128,
validation_data=([val_x_1, val_x_2], val_y_class))
当运行时:
history = model.fit([train_x_1, train_x_2], train_y_class, nb_epoch=300, batch_size=128,
validation_data=([val_x_1, val_x_2], val_y_class))
以下错误发生:
TypeError: unhashable type: 'numpy.ndarray' accours.
所以我检查了
train_x_1
、train_x_2
和 train_y_class
。它们的类型是 <class 'numpy.ndarray'>
。我已经搜索了一个解决方案,所以我尝试将其类型更改为元组,但没有奏效。如果
numpy.ndarray
是不可哈希的,那么 model.fit
接收什么样的输入?训练数据的形状如下:
train_x_1.shape
(2999, 56, 26)
train_x_2.shape
(2999, 56, 26)
train_y_class.shape
(2999, 1)
train_x_1
的一个示例如下:array([[[ 1.62601626e-02, 2.26890756e-01, 1.17764920e-02, ...,
0.00000000e+00, 0.00000000e+00, 0.00000000e+00],
[ 1.62601626e-02, 2.26890756e-01, 1.17764920e-02, ...,
0.00000000e+00, 0.00000000e+00, 0.00000000e+00],
[ 1.62601626e-02, 2.26890756e-01, 1.17764920e-02, ...,
0.00000000e+00, 0.00000000e+00, 0.00000000e+00],
...,
[ 1.62601626e-02, 2.26890756e-01, 1.17764920e-02, ...,
0.00000000e+00, 0.00000000e+00, 0.00000000e+00],
[ 1.62601626e-02, 2.26890756e-01, 1.17764920e-02, ...,
0.00000000e+00, 0.00000000e+00, 0.00000000e+00],
[ 1.62601626e-02, 2.26890756e-01, 1.17764920e-02, ...,
0.00000000e+00, 0.00000000e+00, 0.00000000e+00]],
train_x_1,train_x_2,train_y_class
的个体元素类型也是numpy.ndarray
吗?这可能是问题所在。如果是这种情况,您应该使用张量而不是嵌套的 ndarray。您能否发布这三个变量内容的样本? - Daneel R.Input(shape=(max_sequence,input_dim))
中输入的形状是否与您提供的输入一致?另外,如果train_x_1
和train_x_2
应该成为同一个模型的输入,则最好将它们np.stack在一起,并将堆叠的数组馈送到模型中,而不是将两个张量作为列表的元素传递。train_x_1 = np.reshape(np.array(train_x_1)
为什么要调用np.array? - Daneel R.