Node

[source]

ImageInput class

autokeras.ImageInput(shape=None)

Input node for image data.

The input data should be numpy.ndarray or tf.data.Dataset. The shape of the data should be should be (samples, width, height) or (samples, width, height, channels).


[source]

Input class

autokeras.Input(shape=None)

Input node for tensor data.

The data should be numpy.ndarray or tf.data.Dataset.


[source]

StructuredDataInput class

autokeras.StructuredDataInput(column_names=None, column_types=None, **kwargs)

Input node for structured data.

The input data should be numpy.ndarray, pandas.DataFrame or tensorflow.Dataset. The data should be two-dimensional with numerical or categorical values.

Arguments

  • column_names: A list of strings specifying the names of the columns. The length of the list should be equal to the number of columns of the data. Defaults to None. If None, it will be obtained from the header of the csv file or the pandas.DataFrame.
  • column_types: Dict. The keys are the column names. The values should either be 'numerical' or 'categorical', indicating the type of that column. Defaults to None. If not None, the column_names need to be specified. If None, it will be inferred from the data. A column will be judged as categorical if the number of different values is less than 5% of the number of instances.

[source]

TextInput class

autokeras.TextInput(shape=None)

Input node for text data.

The input data should be numpy.ndarray or tf.data.Dataset. The data should be one-dimensional. Each element in the data should be a string which is a full sentence.