scvi.dataloaders.MappedCollectionDataModule

scvi.dataloaders.MappedCollectionDataModule#

class scvi.dataloaders.MappedCollectionDataModule(collection, batch_key=None, label_key=None, unlabeled_category='Unknown', sample_key=None, batch_size=128, collection_val=None, accelerator='auto', device='auto', shuffle=True, model_name='SCVI', categorical_covariate_keys=None, continuous_covariate_keys=None, **kwargs)[source]#: Bases: LightningDataModule

`batch_labels`
`extra_categorical_covs`
`extra_continuous_covs`
`label_keys`
`labels`
`n_batch`
`n_labels`
`n_obs`
`n_samples`
`n_vars`
`registry`
`sample_keys`
`samples`
`unlabeled_category`	String assigned to unlabeled cells.
`var_names`

`close`()
`inference_dataloader`([shuffle, batch_size, ...])	Dataloader for inference with on_before_batch_transfer applied.
`on_before_batch_transfer`(batch, dataloader_idx)	Convert a lamindb MappedCollection batch to the dictionary scvi-tools expects.
`train_dataloader`()	Create the train data loader.
`val_dataloader`()	Create the validation data loader.

MappedCollectionDataModule.unlabeled_category[source]#: String assigned to unlabeled cells.

MappedCollectionDataModule.inference_dataloader(shuffle=False, batch_size=4096, indices=None, parallel_cpu_count=None)[source]#: Dataloader for inference with on_before_batch_transfer applied.

MappedCollectionDataModule.on_before_batch_transfer(batch, dataloader_idx)[source]#: Convert a lamindb MappedCollection batch to the dictionary scvi-tools expects.

MappedCollectionDataModule.train_dataloader()[source]#

Create the train data loader.

MappedCollectionDataModule.val_dataloader()[source]#

Create the validation data loader.