dtorchtree.datasets

Description

This module contains a collection of datasets that can be used with the dtorchtree module.

The datasets are all simple to load. Here’s an example of loading the dtorchtree.datasets.IrisDataset dataset:

from dtorchtree.datasets import IrisDataset

dataset = IrisDataset()

for sample in dataset:
    print(sample)

The datasets are all subclasses of dtorch.dataset.JDataset.

class IrisDataset(dtorch.dataset.JDataset)

__init__(root: str = './data', split: Tuple[str] | str = 'train', download: bool = False)

Parameters:

root – The root directory to store the dataset.
split – The split to use. Can be either 'train' or 'test'.
download – Whether to download the dataset if it doesn’t exist. Will be ignored if the dataset already exists.

Raises:

ValueError – If split is not 'train' or 'test'.