ValueError：检查输入时出错：预期density_6_input具有3维，但数组的形状

Question

I'm receiving this error from Keras:

ValueError: Error when checking input: expected dense_6_input to have 3 dimensions,      but got array with shape (55, 72)

on

model.fit(X.values, Y.values, nb_epoch=1000, batch_size=16,verbose=0)

This is my code:

    from keras.models import Sequential
    from keras.layers import Dense, Activation

model = Sequential([
    Dense(32, input_shape=X.values.shape),
    Activation('relu'),
    Dense(10),
    Activation('softmax'),
])
model.compile(loss='mse', optimizer='rmsprop')
model.fit(X.values, Y.values, nb_epoch=1000, batch_size=16,verbose=0)

X has a shape of (55, 72)

How can I fix this and what is dense_6_input?

Answer 1

The problem is here:

Dense(32, input_shape=X.values.shape)

Don't set the input_shape to just the shape of the input values array, as the input_shape does not contain the samples dimension. What you want should be:

Dense(32, input_shape=(72,)),

Then you should be able to call fit without issues.