bitorch_engine.layers.qmha.binary.layer.LearnableBias

class bitorch_engine.layers.qmha.binary.layer.LearnableBias(out_chn: int)[source]

A module that introduces a learnable bias term to the input tensor.

This module adds a learnable bias parameter to the input tensor along the channel dimension. The bias parameter is learnable and can be optimized during the training process.

bias

The learnable bias parameter initialized with zeros.

Type:: nn.Parameter

Parameters:: out_chn (int) – The number of output channels. This should match the channel dimension of the input tensor to which the bias will be added.

Methods

`__init__`	Initializes the LearnableBias module.
`forward`	Forward pass of the LearnableBias module.

Attributes

training

__init__(out_chn: int)[source]

Initializes the LearnableBias module.

Parameters:: out_chn (int) – The number of output channels for the bias parameter.

forward(x: Tensor) → Tensor[source]

Forward pass of the LearnableBias module.

Adds the learnable bias to the input tensor.

Parameters:: x (Tensor) – The input tensor to which the learnable bias will be added.
Returns:: The output tensor after adding the learnable bias to the input.
Return type:: Tensor