After training a SAS DLPy model, which method exports it as a SAS ASTORE item for deployment via SAS Model Manager?
Select an answer to reveal the explanation.
Short Explanation
Here's the deal — you save a trained SAS DLPy model for SAS Model Manager deployment by calling model.save_to_astore(), which packages the architecture and weights as a SAS ASTORE item. You don't use the HDF5 export option, a Keras-style serialize method, or PROC ASTORE to create the ASTORE from the model object.
Full Explanation
model.save_to_astore(path) saving as a SAS ASTORE item is correct because SAS DLPy models are saved as SAS ASTORE items using model.save_to_astore(), which packages the model architecture and weights into a portable SAS artifact registerable in SAS Model Manager for deployment and scoring. Saving weights as HDF5 is wrong because DLPy does not natively export to HDF5. Exporting in Keras-compatible format with model.serialize(format='h5') is wrong because model.serialize() is not a DLPy method. Using PROC ASTORE to save an ASTORE is wrong because PROC ASTORE manages existing ASTORE items but does not create them from DLPy model objects.