from keras.wrappers.scikit_learn import KerasClassifier

# create model function to use with KerasClassifier
def create_model(optimizer='adam', neurons=64, dropout_rate=0.25):
    activation='relu'
    #build layers
    model = Sequential()
    model.add(BatchNormalization(axis=-1, momentum=0.99, epsilon=0.001, center=True, scale=True,
                                 beta_initializer='zeros', gamma_initializer='ones', moving_mean_initializer='zeros'))
    model.add(Dense(neurons, activation='relu'))
    model.add(Dropout(dropout_rate))
    model.add(Dense(neurons, activation='relu'))
    model.add(Dropout(dropout_rate))
    model.add(Dense(neurons, activation='relu'))
    model.add(Dropout(dropout_rate))
    model.add(Dense(2, activation='softmax'))
    #compile model
    model.compile(loss='categorical_crossentropy',
                  optimizer=optimizer,
                  metrics=['accuracy'])
    return model

model = KerasClassifier(build_fn=create_model, batch_size=128, epochs=2)

from sklearn.model_selection import RandomizedSearchCV

# create grid of hyperparameters
params = {'neurons':[256, 512],
          'dropout_rate':[0.25, 0.5, 0.75],
          'optimizer':['adam', 'sgd']}

grid = RandomizedSearchCV(estimator=model, param_distributions=params,
                          verbose=2, n_jobs=-1)
grid.fit(X_train, y_train)
grid.best_params_