dlpy.blocks.ResBlock

class dlpy.blocks.ResBlock(kernel_sizes=3, n_filters=(16, 16), strides=None, batch_norm_first=False, conv_short_cut=False)

Bases: object

Base class for the residual blocks.

Parameters:

kernel_sizes : list-of-ints, optional

Specifies the size of the kernels. This assumes the kernels are square.
Default: 3

n_filters : list-of-ints, optional

Specifies the number of filters.
Default: (16, 16)

strides : list-of-ints, optional

Specifies the stride values for the filters

batch_norm_first : bool, optional

Set to True, if the batch normalization comes first
Default: False

conv_short_cut : bool, optional

Set to True, if convolution layer has a short cut
Default: False

Returns:

ResBlock

__init__(kernel_sizes=3, n_filters=(16, 16), strides=None, batch_norm_first=False, conv_short_cut=False)

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

Methods

__init__([kernel_sizes, n_filters, strides, …])	Initialize self.
add_layers()	Add the layers for the block
compile(src_layer[, block_num])	Convert the block structure into DLPy layer definitions.
count_instances()
get_number_of_instances()