Core

Warning

The scvi.core top-level module is PRIVATE. We document here for contributors. Please use the top-level module scvi.models.

Distributions

core.distributions.NegativeBinomial([…])

Negative binomial distribution.

core.distributions.NegativeBinomialMixture(…)

Negative binomial mixture distribution.

core.distributions.ZeroInflatedNegativeBinomial([…])

Zero-inflated negative binomial distribution.

Modules

core.modules.VAE(n_input[, n_batch, …])

Variational auto-encoder model.

core.modules.LDVAE(n_input[, n_batch, …])

Linear-decoded Variational auto-encoder model.

core.modules.TOTALVAE(n_input_genes, …[, …])

Total variational inference for CITE-seq data.

core.modules.SCANVAE(n_input[, n_batch, …])

Single-cell annotation using variational inference.

core.modules.JVAE(dim_input_list, …[, …])

Joint variational auto-encoder for imputing missing genes in spatial data.

core.modules.AutoZIVAE(n_input[, …])

Implementation of the AutoZI model [Clivio19].

core.modules.Classifier(n_input[, n_hidden, …])

Basic fully-connected NN classifier

Data Loaders

core.data_loaders.ScviDataLoader(model, adata)

Scvi Data Loader.

core.data_loaders.TotalDataLoader(model, adata)

Extended data loader for totalVI.

core.data_loaders.AnnotationDataLoader(*args)

Trainers

core.trainers.UnsupervisedTrainer(model, adata)

Class for unsupervised training of an autoencoder.

core.trainers.TotalTrainer(model, dataset[, …])

Unsupervised training for totalVI using variational inference.

core.trainers.SemiSupervisedTrainer(model, adata)

Class for the semi-supervised training of an autoencoder.

core.trainers.ClassifierTrainer(*args[, …])

Class for training a classifier either on the raw data or on top of the latent space of another model.

core.trainers.trainer.Trainer(model, adata)

The abstract Trainer class for training a PyTorch model and monitoring its statistics.

core.trainers.trainer.EarlyStopping([…])

Utilities

core.utils.DifferentialComputation(model_fn, …)

Unified class for differential computation.