hxtorch.spiking.transforms.decode.SumOverTime

class hxtorch.spiking.transforms.decode.SumOverTime(*args, **kwargs)

Bases: torch.nn.modules.module.Module

Simple sum-over-time decoding

__init__(*args, **kwargs) → None

Initializes internal Module state, shared by both nn.Module and ScriptModule.

Methods

forward(input)

Translate an input tensor of shape (batch_size, time_length, population_size) into a tensor of shape (batch_size, population_size), where the time dimension is discarded by computing the sum along the time.

Attributes

forward(input: torch.Tensor) → torch.Tensor

Translate an input tensor of shape (batch_size, time_length, population_size) into a tensor of shape (batch_size, population_size), where the time dimension is discarded by computing the sum along the time. Hence this module performs a ‘sum-over-time’ operation.

Parameters

input – The input tensor to transform. expected shape: (batch_size, time_length, population_size)

Returns

Returns the tensor holding the sum-over-time values.

training: bool