dlpy.layers.Reshape

class dlpy.layers.Reshape(name=None, act='AUTO', fcmp_act=None, width=None, height=None, depth=None, src_layers=None, order=None, **kwargs)

Bases: dlpy.layers.Layer

Reshape layer

Parameters:
name : string, optional

Specifies the name of the layer.

act : string, optional

Specifies the activation function.
Valid Values: AUTO, IDENTITY, LOGISTIC, SIGMOID, TANH, RECTIFIER, RELU, SOFPLUS, ELU, LEAKY, FCMP
Default: AUTO

fcmp_act : string, optional

Specifies the FCMP activation function for the layer.

src_layers : iter-of-Layers, optional

Specifies the layers directed to this layer.

height : int, optional

Specifies the height of the input data. By default the height is determined automatically when the model training begins.

width : int, optional

Specifies the width of the input data. By default the width is determined automatically when the model training begins.

depth : int, optional

Specifies the depth of the feature maps.

src_layers : iter-of-Layers, optional

Specifies the layers directed to this layer.

order : string, optional

Specifies how to reshape the source layer.
Valid Values: AUTO, WHD, WDH, DWH, HWD, DHW, HDW. For AUTO, when the output channel is 1, the reshape order is depth (D), width (W), and height (H). When the output channel is evenly divisible by the input channel, the reshape order is width (W), height (H), and depth (D).

Returns:
Reshape
__init__(name=None, act='AUTO', fcmp_act=None, width=None, height=None, depth=None, src_layers=None, order=None, **kwargs)

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

Methods

__init__([name, act, fcmp_act, width, …]) 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