Skip to content

gemma_cpp

diapason.engine.gemma_cpp

gemma.cpp inference engine backend via pygemma pybind11 bindings.

GemmaCppEngine

Bases: InferenceEngine

gemma.cpp backend via pygemma pybind11 bindings (in-process, CPU).

close()

Unload model and free memory.

prepare(model)

Load model into memory.