blazefl.utils.FilteredDataset

class blazefl.utils.FilteredDataset(indices: list[int], original_data: list, original_targets: list | None = None, transform: Callable | None = None, target_transform: Callable | None = None)

Bases: Dataset

A dataset wrapper that filters and transforms a subset of the original dataset.

This class allows selecting specific data points by their indices and applying optional transformations to the data and targets.

data

The filtered subset of the original dataset.

Type:

list

targets

The filtered subset of targets, if provided.

Type:

list | None

transform

A function to apply transformations to the data.

Type:

Callable | None

target_transform

A function to apply

Type:

Callable | None

transformations to the targets.

__init__(indices: list[int], original_data: list, original_targets: list | None = None, transform: Callable | None = None, target_transform: Callable | None = None) → None

Initialize the FilteredDataset.

Parameters:

• indices (list[int]) – Indices of the data points to include in the dataset.

• original_data (list) – The original dataset.

• original_targets (list | None) – The original targets, if available.

• transform (Callable | None) – Transformation function for the data.

• target_transform (Callable | None) – Transformation function for the targets.

Methods

__init__(indices, original_data[, ...])

Initialize the FilteredDataset.