tqcq/llama.cpp
0
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-07-02 14:15:02 +00:00
Code Issues Packages Projects Releases Wiki Activity

common, server : surface min_keep as its own parameter (#5567)

Browse Source 
* Feature - surface min_keep as its own parameter

* Updated README with min_keep param
This commit is contained in:
Robey Holderith
2024-02-18 11:11:16 -08:00
committed by GitHub
parent c145f8a132
commit 5ee99c32f5
6 changed files with 14 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
common/sampling.h
View File

@ -22,6 +22,7 @@ enum class llama_sampler_type : char {
typedef struct llama_sampling_params {
int32_t n_prev = 64; // number of previous tokens to remember
int32_t n_probs = 0; // if greater than 0, output the probabilities of top n_probs tokens.
int32_t min_keep = 0; // 0 = disabled, otherwise samplers should return at least min_keep tokens
int32_t top_k = 40; // <= 0 to use vocab size
float top_p = 0.95f; // 1.0 = disabled
float min_p = 0.05f; // 0.0 = disabled