Function llama_model_default_params

smallthinker/src/llama-model.cpp:14067–14090

//

llama_model_params llama_model_default_params() {
    llama_model_params result = {
        /*.devices                     =*/ nullptr,
        /*.tensor_buft_overrides       =*/ nullptr,
        /*.n_gpu_layers                =*/ 0,
        /*.split_mode                  =*/ LLAMA_SPLIT_MODE_LAYER,
        /*.main_gpu                    =*/ 0,
        /*.tensor_split                =*/ nullptr,
        /*.progress_callback           =*/ nullptr,
        /*.progress_callback_user_data =*/ nullptr,
        /*.kv_overrides                =*/ nullptr,
        /*.vocab_only                  =*/ false,
        /*.use_mmap                    =*/ true,
        /*.use_mlock                   =*/ false,
        /*.check_tensors               =*/ false
    };

#ifdef GGML_USE_METAL
    // note: we usually have plenty of VRAM, so by default offload all layers to the GPU
    result.n_gpu_layers = 999;
#endif

    return result;
}

const llama_vocab * llama_model_get_vocab(const llama_model * model) {
    return &model->vocab;