jo-mengr's picture
add support for zarr when loading adata objects for tokenizer. Load with anndata.read_zarr().
31c6d20