From 0a1c308d042ef51fdb2246be638c85e95557b9c1 Mon Sep 17 00:00:00 2001
From: CoderRC <108188026+CoderRC@users.noreply.github.com>
Date: Sun, 2 Apr 2023 23:52:27 -0400
Subject: [PATCH] Sync

---
 llama.cpp | 14 ++++++++++++++
 1 file changed, 14 insertions(+)

diff --git a/llama.cpp b/llama.cpp
index 805d6af9e..d0be38438 100644
--- a/llama.cpp
+++ b/llama.cpp
@@ -1242,6 +1242,20 @@ static llama_vocab::id llama_sample_top_p_top_k(
     const auto & logits = lctx.logits;
     const auto * plogits = logits.data() + logits.size() - n_logits;
 
+    if (temp <= 0) {
+        // select the token with the highest logit directly
+        float max_logit = plogits[0];
+        llama_vocab::id max_id = 0;
+
+        for (int i = 1; i < n_logits; ++i) {
+            if (plogits[i] > max_logit) {
+                max_logit = plogits[i];
+                max_id = i;
+            }
+        }
+        return max_id;
+    }
+
     std::vector<std::pair<float, llama_vocab::id>> logits_id;
     logits_id.reserve(n_logits);