tqcq/llama.cpp
0
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-08-19 14:31:06 -04:00
Code Issues Packages Projects Releases Wiki Activity

server : enable KV cache defrag by default (#10233)

Browse Source 
ggml-ci
This commit is contained in:
Georgi Gerganov
2024-11-11 08:38:43 +02:00
committed by GitHub
parent 4b3a9212b6
commit b141e5f6ef
2 changed files with 12 additions and 10 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
common/common.h
View File

@@ -178,7 +178,7 @@ struct common_params {
float yarn_beta_fast = 32.0f; // YaRN low correction dim
float yarn_beta_slow = 1.0f; // YaRN high correction dim
int32_t yarn_orig_ctx = 0; // YaRN original context length
float defrag_thold = -1.0f; // KV cache defragmentation threshold
float defrag_thold = 0.1f; // KV cache defragmentation threshold
struct cpu_params cpuparams;
struct cpu_params cpuparams_batch;