IndexError: list index out of range when using BayesianOptimization

I’m currently having issues trying to BayesianOptimization for the tuning process for a binary classifier. This issue doesn’t appear with RandomSearch and Hyperband. Here’s how the hyper model is constructed:

class MyHyperModel(HyperModel):

    def __init__(self, num_classes):
        self.num_classes = num_classes

    def build(self, hp):

        model = keras.Sequential()

        
        for i in range(hp.Int('num_layers', 2, 20)):

            model.add(keras.layers.Dense(units=hp.Int('units_' + str(i),
                                                min_value=32,
                                                max_value=512, 
                                                step=32),
                                   activation='relu'))

        model.add(keras.layers.Dense(self.num_classes, activation='sigmoid'))

        model.compile(
            optimizer=keras.optimizers.Adam(
                hp.Choice('learning_rate',
                          values=[1e-2, 1e-3, 1e-4])),
            loss='binary_crossentropy',
            metrics=['accuracy'])

        return model


hypermodel = MyHyperModel(num_classes=1)

tuner = BayesianOptimization(
    hypermodel,
    objective='accuracy',
    max_trials=10,
    project_name='test')

tuner.search(X_train.values, y_train.values.flatten(),
             epochs=10,
             validation_data=(X_test.values, y_test.values.flatten()))

The search works well for a while and crashes sometimes with this traceback error. It is worthwhile mentioning that the search even completes without this error but most of the time it crashes to this.

---------------------------------------------------------------------------
IndexError                                Traceback (most recent call last)
<ipython-input-21-d78d7772b203> in <module>()
      1 tuner.search(X_train.values, y_train.values.flatten(),
      2              epochs=10,
----> 3              validation_data=(X_test.values, y_test.values.flatten()))

4 frames
/usr/local/lib/python3.6/dist-packages/kerastuner/engine/hyperparameters.py in cumulative_prob_to_value(prob, hp)
    842         ele_prob = 1 / len(hp.values)
    843         index = math.floor(prob / ele_prob)
--> 844         return hp.values[index]
    845     elif isinstance(hp, (Int, Float)):
    846         sampling = hp.sampling or 'linear'

IndexError: list index out of range

Am I doing something wrong? If yes, how can I correct this?

[omalleyt12]

@PyDataBlog the fix is now merged into master, could you please confirm this fixes the issue you saw, and if not please reopen?

[rafaelstevenson]

Hello, it seems i encountered the same issue but overcame it after updating to 1.0.2 but I have a confusion in using keras tuner for loop for determining hidden layers as search space, if there are anyone willing to help. Thank you https://stackoverflow.com/questions/66783048/keras-tuner-uses-for-i-in-rangehp-intn-layers-1-3-but-does-not-show-agre