From 046c0d77a9f04269171aef16db5a7ad56f075e67 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Stanis=C5=82aw=20Szymczyk?= <sszymczy@gmail.com>
Date: Sat, 14 Dec 2024 12:00:19 +0100
Subject: [PATCH] llama : use zero value of n_swa to distinguish Phi-4 from
 other PHI3 models

---
 src/llama.cpp | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/src/llama.cpp b/src/llama.cpp
index 67d75bb08..7b16daaa5 100644
--- a/src/llama.cpp
+++ b/src/llama.cpp
@@ -5807,7 +5807,7 @@ static void llm_load_hparams(
                     hparams.n_swa = 131072;
                 }
                 bool found_swa = ml.get_key(LLM_KV_ATTENTION_SLIDING_WINDOW, hparams.n_swa, false);
-                if (!found_swa && hparams.n_swa == 0 && model.name != "Phi 4") {
+                if (!found_swa && hparams.n_swa == 0) {
                     throw std::runtime_error("invalid value for sliding_window");
                 }
             } break;
@@ -12840,7 +12840,7 @@ struct llm_build_context {
 
         // KQ_mask (mask for 1 head, it will be broadcasted to all heads)
         struct ggml_tensor * KQ_mask = nullptr;
-        if (model.name == "Phi 4") {
+        if (hparams.n_swa == 0) {
             // Phi-4 doesn't use sliding window attention
             KQ_mask = build_inp_KQ_mask();
         } else {