Scvi

------------------
* update highly variable gene selection to handle sparse matrices `adam`_
* update DE docstrings `pierre`_
* improve posterior save load to also handle subclasses `pierre`_
* Create NB and ZINB distributions with torch and refactor code accordingly `pierre`_
* typos in autozivae `achille`_
* bug in csc sparse matrices in anndata data loader `adam`_

------------------
* handles gene and cell attributes with the same name `han-yuan`_
* fixes anndata overwriting when loading `adam`_, `pierre`_
* formatting in basic tutorial `adam`_

------------------
* updates on TotalVI and LDVAE `adam`_
* fix documentation, compatibility and diverse bugs `adam`_, `pierre`_ `romain`_
* fix for external module on scanpy `galen`_

------------------
* do not automatically upper case genes `adam`_
* AutoZI `oscar`_
* Made the intro tutorial more user friendly `adam`_
* Tests for LDVAE notebook `adam`_
* black codebase `achille`_ `gabriel`_ `adam`_
* fix compatibility issues with sklearn and numba `romain`_
* fix Anndata `francesco-brundu`_
* docstring, totalVI, totalVI notebook and CITE-seq data `adam`_
* fix type `eduardo-beltrame`_
* fixing installation guide `jeff`_
* improved error message for dispersion `stephen-flemming`_

------------------

* docstring `achille`_
* differential expression `oscar`_ `pierre`_

------------------

* gimVI `achille`_
* synthetic correlated datasets, fixed bug in marginal log likelihood `oscar`_
* autotune, dataset enhancements `gabriel`_
* documentation `jeff`_
* more consistent posterior API, docstring, validation set `adam`_
* fix anndataset `michael-raevsky`_
* linearly decoded VAE `valentine-svensson`_
* support for scanpy, fixed bugs, dataset enhancements `achille`_
* fix filtering bug, synthetic correlated datasets, docstring, differential expression `pierre`_
* better docstring `jamie-morton`_
* classifier based on library size for doublet detection `david-kelley`_