From f430ead1c28c198fd445b6d7027a01e8e228a024 Mon Sep 17 00:00:00 2001
From: vansangpfiev <vansangpfiev@gmail.com>
Date: Tue, 22 Oct 2024 11:04:43 +0700
Subject: [PATCH] fix: warm-up (#259)

* fix: warm-up

* fix: n_predict 10

* Revert "fix: n_predict 10"

This reverts commit d08e95289773753c23c25630efa1942754e0c3ac.

* revert: CI

---------

Co-authored-by: vansangpfiev <sang@jan.ai>
---
 src/llama_engine.cc | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/src/llama_engine.cc b/src/llama_engine.cc
index bfcdb32..06f2874 100644
--- a/src/llama_engine.cc
+++ b/src/llama_engine.cc
@@ -978,6 +978,8 @@ void LlamaEngine::WarmUpModel(const std::string& model_id) {
     pseudo["prompt"] = "Hello";
     pseudo["n_predict"] = 2;
     pseudo["stream"] = false;
+    pseudo["cache_prompt"] = server_map_[model_id].caching_enabled;
+    pseudo["n_keep"] = 0;
     const int task_id =
         si->second.ctx.RequestCompletion(pseudo, false, false, -1);
     TaskResult result = si->second.ctx.NextResult(task_id);