scvi.dataloaders.SemiSupervisedDataLoader#

class scvi.dataloaders.SemiSupervisedDataLoader(adata_manager, n_samples_per_label=None, indices=None, shuffle=False, batch_size=128, data_and_attributes=None, drop_last=False, **data_loader_kwargs)[source]#

Bases: ConcatDataLoader

DataLoader that supports semisupervised training.

Parameters:
adata_manager AnnDataManager

AnnDataManager object that has been created via setup_anndata.

n_samples_per_label int | None (default: None)

Number of subsamples for each label class to sample per epoch. By default, there is no label subsampling.

indices list[int] | None (default: None)

The indices of the observations in the adata to load

shuffle bool (default: False)

Whether the data should be shuffled

batch_size int (default: 128)

minibatch size to load each iteration

data_and_attributes dict | None (default: None)

Dictionary with keys representing keys in data registry (adata_manager.data_registry) and value equal to desired numpy loading type (later made into torch tensor). If None, defaults to all registered data.

data_loader_kwargs

Keyword arguments for DataLoader

Attributes table#

Methods table#

resample_labels()

Resamples the labeled data.

subsample_labels()

Subsamples each label class by taking up to n_samples_per_label samples per class.

Attributes#

SemiSupervisedDataLoader.dataset#
SemiSupervisedDataLoader.batch_size#
SemiSupervisedDataLoader.num_workers#
SemiSupervisedDataLoader.pin_memory#
SemiSupervisedDataLoader.drop_last#
SemiSupervisedDataLoader.timeout#
SemiSupervisedDataLoader.sampler#
SemiSupervisedDataLoader.pin_memory_device#
SemiSupervisedDataLoader.prefetch_factor#

Methods#

SemiSupervisedDataLoader.resample_labels()[source]#

Resamples the labeled data.

SemiSupervisedDataLoader.subsample_labels()[source]#

Subsamples each label class by taking up to n_samples_per_label samples per class.