happyz
/
llama.cpp
mirror of https://github.com/ggerganov/llama.cpp.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

server: save generated text for the /slots endpoint (for LLAMA_SERVER_SLOTS_DEBUG=1) (#19622) 

* save generated text for the /slots endpoint

* update debug_generated_text only when LLAMA_SERVER_SLOTS_DEBUG > 0

* Apply suggestions from code review

---------

Co-authored-by: Matteo <matteo@matteo>
Co-authored-by: Xuan-Son Nguyen <thichthat@gmail.com>
This commit is contained in:
matteo 2026-02-18 18:53:37 +01:00 committed by GitHub
parent eeef3cfced
commit b55dcdef5d
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 8 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

9
tools/server/server-context.cpp
View File

@ -77,6 +77,7 @@ struct server_slot {
size_t last_nl_pos = 0;
std::string generated_text;
std::string debug_generated_text;
llama_tokens generated_tokens;
// idx of draft tokens in the main batch
@ -425,7 +426,7 @@ struct server_slot {
if (!only_metrics) {
res["prompt"] = ptask->tokens.detokenize(ctx, true);
res["generated"] = generated_text;
res["generated"] = generated_text.empty() ? debug_generated_text : generated_text;
}
}
@ -1442,6 +1443,12 @@ private:
res->id_slot = slot.id;
res->index = slot.task->index;
// keep copy of last generated text for debugging purposes
if (slots_debug) {
slot.debug_generated_text = slot.generated_text;
}
// in stream mode, content and tokens are already in last partial chunk
if (slot.task->params.stream) {
res->content = "";