menu
llama_library package
documentation
io/models/context_params.dart
LLamaContextParams
offloadKqv property
offloadKqv property
dark_mode
light_mode
offloadKqv
property
bool
offloadKqv
getter/setter pair
Whether to offload the KQV ops (including the KV cache) to GPU
Implementation
bool offloadKqv = true;
llama_library package
documentation
io/models/context_params
LLamaContextParams
offloadKqv property
LLamaContextParams class