tfep.io.dataset.dict.DictDataset

class tfep.io.dataset.dict.DictDataset(tensor_dict: dict[str, Sequence])

Bases: Dataset

Utility class to create a map-style PyTorch Datasets from a dictionary of tensors.

The class automatically converts non-tensor dictionary values into tensors.

Examples

>>> import torch
>>> data = {'a': torch.tensor([1.0, 2.0]), 'b': [3, 4]}
>>> dict_dataset = DictDataset(data)
>>> dict_dataset[1]
{'a': tensor(2.), 'b': tensor(4)}

__init__(tensor_dict: dict[str, Sequence])

Constructor.

Parameters:

tensor_dict (dict[str, torch.Tensor]) – A dictionary of named tensors.

Methods

__init__(tensor_dict)

Constructor.