server: add checkpoint tolerance and fix grammar_trigger init (#1346)

Co-authored-by: firecoperana <firecoperana>
This commit is contained in:
firecoperana
2026-03-02 00:45:32 -06:00
committed by GitHub
parent a568e12c8f
commit 8f9e19d57c
4 changed files with 27 additions and 4 deletions

View File

@@ -105,6 +105,7 @@ struct server_slot {
void prompt_load(server_prompt_cache& prompt_cache, const server_tokens& tokens);
size_t checkpoint_pos = 0;
bool do_checkpoint = false;
// sampling
llama_token sampled; // in speculative mode, this is the last accepted token