Question

I recently wrote an LSTM model to predict sequence:

    ############### BUILD MODEL ###############

''' HERE WE ARE CREATING THE LSTM MODEL '''

model = Sequential()
model.add(LSTM(128, input_shape=(X.shape[1:]), activation='relu', return_sequences=True))

model.add(Dropout(0.2))

model.add(LSTM(128,input_shape=(X.shape[1:]), activation='relu'))
model.add(Dropout(0.1))

model.add(Dense(32, activation='relu'))
model.add(Dropout(0.2))

model.add(Dense(10, activation='softmax'))

# In[8]:


'''HERE WE ARE CREATING AN OPTIMIZER AND THEN TRAINING OUR MODEL'''

opt = tf.keras.optimizers.Adam(lr=0.001, decay=1e-6)

model.compile(
    loss='sparse_categorical_crossentropy',
    optimizer=opt,
    metrics=['accuracy'],
)

model.fit(X, Y, batch_size=10, epochs=1)

where np.shape(X) = (237, 30, 3) and np.shape(Y) = (237, 3). But while fitting this data to the model it is returning an error:

ValueError: Error when checking target: expected dense_1 to have shape (1,) but got array with shape (3,)

What is wrong in this code?

Answer 1

It seems you have one-hot encoded the labels. Either don't one-hot encode them (i.e. let them be sparse labels) and use sparse_categorical_crossentropy as the loss function or alternatively, one-hot encode them and use categorical_crossentropy as the loss function.

As a side note: are you sure it is a classification task? Since you are using a softmax layer as the last layer with 10 units but you mentioned that the labels have 3 classes?!

Answer 2

如果数据形状是这样的：np.shape（X）=（237，30，3）和np.shape（Y）=（237，3）试试这个网络。

model = Sequential()
model.add(LSTM(32, input_shape=(X.shape[1:]), activation='relu', return_sequences=True))
model.add(Dropout(0.2))
model.add(LSTM(32, activation='relu'))
model.add(Dropout(0.1))
model.add(Dense(32, activation='relu'))
model.add(Dropout(0.2))
model.add(Dense(3, activation='softmax'))

opt = tf.keras.optimizers.Adam(lr=0.001, decay=1e-6)

model.compile(
    loss='categorical_crossentropy',
    optimizer=opt,
    metrics=['accuracy'],
)

model.fit(X, Y, batch_size=10, epochs=1)

由于使用的是Keras Sequential Api，因此您不必为第二个LSTM层指定输入，并且在最终的Dense层，您的“ Y”形状为（237，3），因此最终的Dense层也应为也给出3。

如果每个样本中输入中的“ 3”（要素数量）恒定，则还可以为第一个LSTM层输入input_shape =（None，3）。

How to train an LSTM model on multi-dimentional array?

2 个答案: