Attention

June 2024 Status Update: Removing DataPipes and DataLoader V2

We are re-focusing the torchdata repo to be an iterative enhancement of torch.utils.data.DataLoader. We do not plan on continuing development or maintaining the [DataPipes] and [DataLoaderV2] solutions, and they will be removed from the torchdata repo. We’ll also be revisiting the DataPipes references in pytorch/pytorch. In release torchdata==0.8.0 (July 2024) they will be marked as deprecated, and in 0.9.0 (Oct 2024) they will be deleted. Existing users are advised to pin to torchdata==0.8.0 or an older version until they are able to migrate away. Subsequent releases will not include DataPipes or DataLoaderV2. Please reach out if you suggestions or comments (please use this issue for feedback)

ShardExpander

class torchdata.datapipes.iter.ShardExpander(source_datapipe: IterDataPipe[str])

Expands incoming shard strings into shards.

Sharded data files are named using shell-like brace notation. For example, an ImageNet dataset sharded into 1200 shards and stored on a web server might be named imagenet-{000000..001199}.tar.

Note that shard names can be expanded without any server transactions; this makes shard_expand reproducible and storage system independent (unlike :class .FileLister etc.).

Parameters:

source_datapipe – a DataPipe yielding a stream of pairs

Returns:

a DataPipe yielding a stream of expanded pathnames.

Example

>>> from torchdata.datapipes.iter import IterableWrapper
>>> source_dp = IterableWrapper(["ds-{00..05}.tar"])
>>> expand_dp = source_dp.shard_expand()
>>> list(expand_dp)
['ds-00.tar', 'ds-01.tar', 'ds-02.tar', 'ds-03.tar', 'ds-04.tar', 'ds-05.tar']
>>> source_dp = IterableWrapper(["imgs_{00..05}.tar", "labels_{00..05}.tar"])
>>> expand_dp = source_dp.shard_expand()
>>> list(expand_dp)
['imgs_00.tar', 'imgs_01.tar', 'imgs_02.tar', 'labels_00.tar', 'labels_01.tar', 'labels_02.tar']