Closed ViktorKaz closed 6 years ago
This was caused by a circular reference when calling predict_class()
from the mlaut.estimators.nn_estimators.OverwrittenSequentialClassifier
Fixed by calling predict()
and calculating the max:
predictions = Sequential.predict(self, X_test, batch_size=batch_size, verbose=verbose)
return predictions.argmax(axis=1)
There is a circular reference error when trying to load saved keras models from disk.