LLamaSharp/docs/xmldocs/llama.native.llamacontextparams.md

• LLamaContextParams
• • Fields
  • • n_ctx
    • n_gpu_layers
    • seed
    • f16_kv
    • logits_all
    • vocab_only
    • use_mmap
    • use_mlock
    • embedding
    • progress_callback
    • progress_callback_user_data

LLamaContextParams

Namespace: LLama.Native

public struct LLamaContextParams

Inheritance Object → ValueType → LLamaContextParams

Fields

n_ctx

text context

public int n_ctx;

n_gpu_layers

number of layers to store in VRAM

public int n_gpu_layers;

seed

RNG seed, -1 for random

public int seed;

f16_kv

use fp16 for KV cache

public bool f16_kv;

logits_all

the llama_eval() call computes all logits, not just the last one

public bool logits_all;

vocab_only

only load the vocabulary, no weights

public bool vocab_only;

use_mmap

use mmap if possible

public bool use_mmap;

use_mlock

force system to keep model in RAM

public bool use_mlock;

embedding

embedding mode only

public bool embedding;

progress_callback

called with a progress value between 0 and 1, pass NULL to disable

public IntPtr progress_callback;

progress_callback_user_data

context pointer passed to the progress callback

public IntPtr progress_callback_user_data;