Keras : problem in fitting model

Question

I am trying to fit a keras classifier on a data matrix X_train

dummy_y = np_utils.to_categorical(y_train)
dummy_y_val = np_utils.to_categorical(y_val)

def baseline_model(optimizer='rmsprop', init='normal', dropout_rate =0.0):
    model = Sequential()
    model.add(Dense(100, input_dim = X_train.shape[1], activation='relu', init=init))
    model.add(Dropout(dropout_rate))
    model.add(Dense(50, activation = 'relu', init = init))
    model.add(Dropout(dropout_rate))
    model.add(Dense(15, activation = 'sigmoid', init = init))
    model.compile(loss = 'binary_crossentropy', optimizer = optimizer, metrics =['accuracy'])
    return model

   estimator = KerasClassifier(build_fn=baseline_model, nb_epoch=200, batch_size=5, verbose=0)
   model = baseline_model()

model.fit(X_train, dummy_y, batch_size=32,  verbose=1, callbacks=None, \
     validation_data=(X_val, dummy_y_val), shuffle=True, class_weight=None, \
    sample_weight=None, initial_epoch=0)

That's what I get as an error :

Train on 582 samples, validate on 290 samples Epoch 1/10

ValueErrorTraceback (most recent call last) in () 1 model = baseline_model() ----> 2 model.fit(X_train, dummy_y, batch_size=32, verbose=1,

....

ValueError: Cannot feed value of shape (582, 32) for Tensor u'dense_input_28:0', which has shape '(?, 18760)'

NOTE : X_train.shape = (580,18760) and y_train has 15 classes

Answer 1

The problem was : X_train and y_train were not numpy array. I transformed them to numpy arrays and it solved the problem.