Yes, @FrozenGene is correct. At the moment, only offload is suported (via pipeline attribute). Running entire models is on our roadmap, but there aren’t any specific dates when this functionality will be available.
Yes, @FrozenGene is correct. At the moment, only offload is suported (via pipeline attribute). Running entire models is on our roadmap, but there aren’t any specific dates when this functionality will be available.