mindspore.dataset.dataloader

This module provides iterators for loading datasets. It supports loading both Map Style and Iterable Style datasets, and offers multi-process concurrent loading.

Data Loader

mindspore.dataset.dataloader.DataLoader

Data loader provides an iterator over the given dataset.

Datasets

`mindspore.dataset.dataloader.Dataset`	Base class for implementing all datasets.
`mindspore.dataset.dataloader.IterableDataset`	Base class for implementing iterable datasets.
`mindspore.dataset.dataloader.TensorDataset`	Dataset that defined by a collection of `mindspore.Tensor` .

Samplers

`mindspore.dataset.dataloader.Sampler`	Base Class of the Sampler
`mindspore.dataset.dataloader.SequentialSampler`	Samples the dataset elements sequentially.
`mindspore.dataset.dataloader.RandomSampler`	Samples the dataset elements randomly.
`mindspore.dataset.dataloader.BatchSampler`	Sampler that yields a mini-batch of indices each time.
`mindspore.dataset.dataloader.DistributedSampler`	A sampler that partitioning datasets for distributed training.

Collate Functions

`mindspore.dataset.dataloader.default_collate`	Default function for concatenating a batch of data along the first dimension when batching is enabled in `DataLoader`.
`mindspore.dataset.dataloader.default_convert`	Default function for converting each NumPy array element into a `mindspore.Tensor` when batching is disabled in `DataLoader`.
`mindspore.dataset.dataloader._utils.collate.collate`	Collate the input batch of data by the appropriate function for each element type selected from the type to collate function mapping defined in collate_fn_map.

Utilities

mindspore.dataset.dataloader.get_worker_info

Get the information about the current DataLoader worker process.