converter

converter

Functions

Name	Description
isSNNLayer	Checks if a layer is an instance of a Spiking Neural Network (SNN) layer.
pad_with	Pads an array with a specified value.
weight_quantization	Applies weight quantization to the input.

converter.isSNNLayer(layer)

Checks if a layer is an instance of a Spiking Neural Network (SNN) layer.

Name	Type	Description	Default
`layer`	object	The layer to check.	required

Type	Description
bool	True if the layer is an instance of a SNN layer, False otherwise.

>>> from norse.torch.module.lif import LIFCell
>>> layer = LIFCell()
>>> isSNNLayer(layer)
True

converter.pad_with(vector, pad_width, iaxis, kwargs)

Pads an array with a specified value.

Name	Type	Description	Default
`vector`	numpy.ndarray	Input array.	required
`pad_width`	int or sequence of ints	Number of values padded to the edges of each axis.	required
`iaxis`	int	An indicator of the current axis.	required
`kwargs`	dict	Optional keyword arguments.	required

Type	Description
numpy.ndarray	The padded array.

>>> import numpy as np
>>> a = np.array([1, 2, 3, 4, 5])
>>> np.pad(a, 2, pad_with, padder=2)
array([2, 2, 1, 2, 3, 4, 5, 2, 2])

converter.weight_quantization(b)

Applies weight quantization to the input.

Name	Type	Description	Default
`b`	int	The number of bits to use for the quantization.	required

Type	Description
function	A function that applies weight quantization to its input.

>>> weight_quantization_func = weight_quantization(8)
>>> weight_quantization_func(some_input)