CS348Project/llama.cpp
5
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-10-28 08:31:25 +00:00
Code Issues Packages Projects Releases Wiki Activity

server : adjust prompt similarity thold + add logs (#15913)

Browse Source 
ggml-ci
This commit is contained in:
Georgi Gerganov
2025-09-12 17:02:55 +03:00
committed by GitHub
parent 304ac5693d
commit f088b6a84f
2 changed files with 3 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
common/common.h
View File

@@ -452,7 +452,7 @@ struct common_params {
std::string slot_save_path; std::string slot_save_path;
float slot_prompt_similarity = 0.5f; float slot_prompt_similarity = 0.1f;
// batched-bench params // batched-bench params
bool is_pp_shared = false; bool is_pp_shared = false;

4
tools/server/server.cpp
View File

@@ -2372,7 +2372,7 @@ struct server_context {
} }
if (ret != nullptr) { if (ret != nullptr) {
SLT_DBG(*ret, "selected slot by lcs similarity, lcs_len = %d, similarity = %f\n", lcs_len, similarity); SLT_INF(*ret, "selected slot by lcs similarity, lcs_len = %d, similarity = %.3f (> %.3f thold)\n", lcs_len, similarity, slot_prompt_similarity);
} }
} }
@@ -2394,7 +2394,7 @@ struct server_context {
} }
if (ret != nullptr) { if (ret != nullptr) {
SLT_DBG(*ret, "selected slot by lru, t_last = %" PRId64 "\n", t_last); SLT_INF(*ret, "selected slot by LRU, t_last = %" PRId64 "\n", t_last);
} }
} }