nyoka AttributeError: The layer has never been called and thus has no defined input shape

Question

I'm trying to output a trained Tensorflow 2.0 model to PMML using the nyoka package. When I do so, it errors out. The problem seems to be different from that in this answer, even though the error is the same, because my model does not have a complicated creation function and does, in fact, train appropriately and transform appropriately.

def create_and_train(x_training,y_training,n_cols_in,modelparams):
    layers = [tf.keras.layers.Dense(n_cols_in,activation="relu"),
    tf.keras.layers.Dropout(.5)]
    for param in modelparams:
        layers.extend([tf.keras.layers.Dense(param,activation="sigmoid"),tf.keras.layers.Dropout(.5)])
    layers.append(tf.keras.layers.Dense(1,activation="sigmoid"))
    model = tf.keras.models.Sequential(layers)
    model.compile(optimizer='adam', loss='binary_crossentropy', metrics=[tf.keras.metrics.AUC()])

    model.fit(x_training, y_training, epochs = epochs)
    with open("NN"+"_".join([str(m) for m in modelparams])+".pmml","w") as pmml_file:
        pmml = KerasToPmml(model)
        pmml.export(pmml_file)

Instead of a PMML file, I'm getting

AttributeError: The layer has never been called and thus has no defined input shape.

score 0 · Accepted Answer · answered Jul 14 '19 at 09:43

0

This is Tensorflow's error. If you can print input_shape and output_shape or weights for each layer then you will be able to export it using Nyoka also.

answered Jul 14 '19 at 09:43

Nirmal

116
1
4

nyoka AttributeError: The layer has never been called and thus has no defined input shape

1 Answers1