scvi-tools, the deep generative single-cell embedding used as a learned-representation comparator.

scvi

scvi-tools learns a probabilistic latent representation of single-cell data. It is a companion to the single-cell vertical rather than part of it: a comparison method that is not topological, which is what makes it useful when the question is whether the topological scaffold earns its place.

Resolves from conda-forge, locked green.