multiVI and totalVI modal integration question


Recently, I am using scvi-tools to generate multi-modal integrated datasets.

I have a question about the MULTIVI.get_latent_representation and TOTALVI.get_latent_representation function.

I followed the user guide to do the integration.

My question is for the TOTALVI.get_latent_representation function, I can get the exactly same number of total cells from the latent space matrix. But for the MULTIVI.get_latent_representation, the latent space matrix includes 3 times the total cells of my original datasets ( original 1000 cells, latent space 3000 cells). I think that may be because for the multiVI we concatenate 3 anndata
adata_mvi =, adata_rna, adata_atac).
If I want to get the multimodal integrated dataset from multiVI, should I only use the first 1000 rows as the integrated result from RNA and ATAC modality?