Struct llama_model_params 

#[repr(C)]
pub struct llama_model_params {

    pub devices: *const c_char,
    pub n_gpu_layers: i32,
    pub mla: i32,
    pub split_mode: llama_split_mode,
    pub main_gpu: i32,
    pub max_gpu: i32,
    pub ncmoe: i32,
    pub type_k: ggml_type,
    pub type_v: ggml_type,
    pub idx_type_k: ggml_type,
    pub max_ctx_size: u32,
    pub n_seq_max: i32,
    pub n_ubatch: i32,
    pub amb: i32,
    pub fit_margin: i32,
    pub fit: bool,
    pub worst_graph_tokens: i32,
    pub type_k_first: ggml_type,
    pub type_k_last: ggml_type,
    pub type_v_first: ggml_type,
    pub type_v_last: ggml_type,
    pub n_k_first: i32,
    pub n_k_last: i32,
    pub n_v_first: i32,
    pub n_v_last: i32,
    pub extra_output_type: ggml_type,
    pub tensor_split: *const f32,
    pub fit_margin_array: *const c_int,
    pub rpc_servers: *const c_char,
    pub progress_callback: llama_progress_callback,
    pub progress_callback_user_data: *mut c_void,
    pub kv_overrides: *const llama_model_kv_override,
    pub tensor_buft_overrides: *const llama_model_tensor_buft_override,
    pub vocab_only: bool,
    pub use_mmap: bool,
    pub use_mlock: bool,
    pub check_tensors: bool,
    pub repack_tensors: bool,
    pub use_thp: bool,
    pub validate_quants: bool,
    pub merge_qkv: bool,
    pub merge_up_gate_exps: bool,
    pub mtp: bool,
    pub dry_run: bool,
    pub flash_attn: bool,
    pub defer_experts: bool,
}

Fields

devices: *const c_char

n_gpu_layers: i32

mla: i32

split_mode: llama_split_mode

main_gpu: i32

max_gpu: i32

ncmoe: i32

type_k: ggml_type

type_v: ggml_type

idx_type_k: ggml_type

max_ctx_size: u32

n_seq_max: i32

n_ubatch: i32

amb: i32

fit_margin: i32

fit: bool

worst_graph_tokens: i32

type_k_first: ggml_type

type_k_last: ggml_type

type_v_first: ggml_type

type_v_last: ggml_type

n_k_first: i32

n_k_last: i32

n_v_first: i32

n_v_last: i32

extra_output_type: ggml_type

tensor_split: *const f32

fit_margin_array: *const c_int

rpc_servers: *const c_char

progress_callback: llama_progress_callback

progress_callback_user_data: *mut c_void

kv_overrides: *const llama_model_kv_override

tensor_buft_overrides: *const llama_model_tensor_buft_override

vocab_only: bool

use_mmap: bool

use_mlock: bool

check_tensors: bool

repack_tensors: bool

use_thp: bool

validate_quants: bool

merge_qkv: bool

merge_up_gate_exps: bool

mtp: bool

dry_run: bool

flash_attn: bool

defer_experts: bool

Trait Implementations

impl Clone for llama_model_params

fn clone(&self) -> llama_model_params

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Copy for llama_model_params

impl Debug for llama_model_params

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl PartialEq for llama_model_params

fn eq(&self, other: &llama_model_params) -> bool

Equality operator ==.

fn ne(&self, other: &Rhs) -> bool

Inequality operator !=.

impl StructuralPartialEq for llama_model_params