Dataloader

Available dataloaders

[source]

Dataset

kipoi.data.Dataset()

An abstract class representing a Dataset.

All other datasets should subclass it. All subclasses should override __len__, that provides the size of the dataset, and __getitem__, supporting integer indexing in range from 0 to len(self) exclusive.


[source]

BatchDataset

kipoi.data.BatchDataset()

An abstract class representing a BatchDataset.


[source]

SampleIterator

kipoi.data.SampleIterator()

[source]

BatchIterator

kipoi.data.BatchIterator()

get_dataloader_factory

get_dataloader_factory(dataloader, source='kipoi')

PreloadedDataset

PreloadedDataset(self)

Generated by supplying a function returning the full dataset.

The full dataset is a nested (list/dict) python structure of numpy arrays with the same first axis dimension.


SampleGenerator

SampleGenerator(self)

Transform a generator of samples into SampleIterator


BatchGenerator

BatchGenerator(self)

Transform a generator of batches into BatchIterator