It's still lacking llama.cpp support, and the work on that isn't moving very fast either. Looks more like a general community issue, where this model isn't drawing much interest.