Hello everyone,

I am trying to deconvolute the spatial data using single cell data annotation using cell2location. My first step is all fine. However, when I am loading the model again, it gives me the following error.

Code used :

adata_ref = adata_ref[:, selected].copy()
cell2location.models.RegressionModel.setup_anndata(adata=adata_ref,
                   # 10X reaction / sample / batch
                   batch_key='orig.ident',
                   # cell type, covariate used for constructing signatures
                   labels_key='celltype_sub',
                   # multiplicative technical effects (platform, 3' vs 5', donor effect)
                  )

# create the regression model
from cell2location.models import RegressionModel

mod = RegressionModel(adata_ref)

mod.train(max_epochs=300, use_gpu=True)
mod.save(f"{ref_run_name}", overwrite=True)
adata_file = f"{ref_run_name}/sc.h5ad"
adata_ref.write(adata_file)

# load the model
mod = cell2location.models.Cell2location.load(f"{run_name}", adata_ref)

Error I get :

ValueError                                Traceback (most recent call last)
Cell In[8], line 1
----> 1 mod = cell2location.models.Cell2location.load(f"{run_name}", adata_ref)
File ~/miniconda3/envs/cell2loc_env/lib/python3.9/site-packages/scvi/model/base/_base_model.py:702, in BaseModelClass.load(cls, dir_path, adata, use_gpu, accelerator, device, prefix, backup_url)
    700 registry = attr_dict.pop("registry_")
    701 if _MODEL_NAME_KEY in registry and registry[_MODEL_NAME_KEY] != cls.__name__:
--> 702     raise ValueError(
    703         "It appears you are loading a model from a different class."
    704     )
    706 if _SETUP_ARGS_KEY not in registry:
    707     raise ValueError(
    708         "Saved model does not contain original setup inputs. "
    709         "Cannot load the original setup."
    710     )
ValueError: It appears you are loading a model from a different class.

By any chance, has anyone else encountered same error, can someone please suggest the issue. I am mostly sure with the previous code.