kaira.models.binary.soft_bit_thresholding.WeightedThresholder

Inheritance diagram of WeightedThresholder — Inheritance diagram for WeightedThresholder

class kaira.models.binary.soft_bit_thresholding.WeightedThresholder(weights: Tensor | List[float] | float, threshold: float = 0.5, input_type: InputType = InputType.PROBABILITY, normalize_weights: bool = False, *args: Any, **kwargs: Any)[source]

Bases: SoftBitThresholder

Thresholder that applies weights to input values before thresholding.

This thresholder allows applying non-uniform weights to different parts of the input tensor, which is useful for systems where some bits are more reliable or important than others.

Example

With weights=[1.0, 0.8, 0.5], threshold=0.6: Input [0.7, 0.7, 0.7] becomes [1.0, 0.0, 0.0] after weighting.

Methods

`__init__`	Initialize the weighted thresholder.
`forward`	Apply weighted thresholding to convert soft bit values to hard decisions.
`to`	Move the model to the specified device.

__init__(weights: Tensor | List[float] | float, threshold: float = 0.5, input_type: InputType = InputType.PROBABILITY, normalize_weights: bool = False, *args: Any, **kwargs: Any)[source]

Initialize the weighted thresholder.

Parameters:

weights – Weights to apply to input values. Can be a tensor, list, or scalar.
threshold – Threshold value to apply after weighting.
input_type – Type of soft input values.
normalize_weights – If True, weights are normalized to sum to 1.0.
*args – Variable positional arguments passed to the base class.
**kwargs – Variable keyword arguments passed to the base class.

forward(x: Tensor, *args: Any, **kwargs: Any) → Tensor[source]

Apply weighted thresholding to convert soft bit values to hard decisions.

Parameters:

x – Input tensor of soft bit values.
*args – Additional positional arguments (unused).
**kwargs – Additional keyword arguments (unused).

Returns:

Tensor of hard bit decisions (0.0 or 1.0).

to(device: str | device, *args, **kwargs) → SoftBitThresholder

Move the model to the specified device.

Parameters:

device – The device to move the model to.
*args – Additional positional arguments for nn.Module.to().
**kwargs – Additional keyword arguments for nn.Module.to().

Returns:

Self for method chaining.