In mnist.py example at the line 187 there are three useless lines
(
Keras expects DenseVectors.
dense_transformer = DenseTransformer(input_col="features_normalized", output_col="features_normalized_dense")
dataset_train = dense_transformer.transform(dataset_train)
dataset_test = dense_transformer.transform(dataset_test)
)
as "matrix" is already defined and as "features_normalized_dense" is unused.
In mnist.py example at the line 187 there are three useless lines (
Keras expects DenseVectors.
dense_transformer = DenseTransformer(input_col="features_normalized", output_col="features_normalized_dense") dataset_train = dense_transformer.transform(dataset_train) dataset_test = dense_transformer.transform(dataset_test) ) as "matrix" is already defined and as "features_normalized_dense" is unused.
mnist.py.zip