dlpy.layers.InputLayer

class dlpy.layers.InputLayer(n_channels=None, width=None, height=None, name=None, nominals=None, std=None, scale=None, offsets=None, dropout=None, random_flip=None, random_crop=None, random_mutation=None, norm_stds=None, **kwargs)

Bases: dlpy.layers.Layer

Input layer

Parameters:

name : string

Specifies the name of the input layer.

nominals : string-list, optional

Specifies the nominal input variables to use in the analysis.

std : string, optional

Specifies how to standardize the variables in the input layer.
Valid Values: MIDRANGE, NONE, STD

n_channels : int, optional for rnn required for cnn

Specifies the depth of the input data, used if data is image.

width : int, optional for rnn required for cnn

Specifies the width of the input data, used if data is image.

height : int, optional

Specifies the height of the input data, used if data is image.
Note: Required for CNN.

scale : float, optional,

Specifies the scale to be used to scale the input data.

offsets : int-list, optional

Specifies the values to be subtracted from the pixel values of the input data, used if the data is image.

dropout : float, optional

Specifies the dropout rate.

random_flip : int, optional

Specifies the type of the random flip to be applied to the input data, used if data is image.
Valid Values: NONE, H, V, HV
Default: NONE

random_crop : int, optional

Specifies the type of the random crop to be applied to the input data, used if data is image.
Valid Values: NONE, UNIQUE
Default: NONE

random_mutation : int, optional

Specifies the type of the random mutation to be applied to the input data, used if data is image.
Valid Values: NONE, RANDOM
Default: NONE

norm_stds : float-list, optional

Specifies a standard deviation for each channel in the input data. The final input data is normalized with specified means and standard deviations.
Default: NONE

Returns:

InputLayer

__init__(n_channels=None, width=None, height=None, name=None, nominals=None, std=None, scale=None, offsets=None, dropout=None, random_flip=None, random_crop=None, random_mutation=None, norm_stds=None, **kwargs)

Initialize self. See help(type(self)) for accurate signature.

Methods

__init__([n_channels, width, height, name, …])	Initialize self.
count_instances()
format_name([block_num, local_count])	Format the name of the layer
get_number_of_instances()
to_model_params()	Convert the model configuration to CAS action parameters