From fb05f80f89fbd38636c06ea7e0d4dd2217bb03d3 Mon Sep 17 00:00:00 2001
From: Lizonghang <870644199@qq.com>
Date: Thu, 23 Jan 2025 16:58:25 +0400
Subject: [PATCH] remove token_embd from metal mem

---
 src/llama.cpp | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/src/llama.cpp b/src/llama.cpp
index de73b7d2..746e847b 100644
--- a/src/llama.cpp
+++ b/src/llama.cpp
@@ -5260,7 +5260,7 @@ struct llama_model_loader {
                 try {
                     const char * tname = ggml_get_name(tensor);
                     if (keep_only_inp_out && !(
-                            strcmp(tname, "token_embd.weight") == 0 || 
+                            // strcmp(tname, "token_embd.weight") == 0 || // lookup table is used so we do not need to keep it in metal memory
                             strcmp(tname, "output_norm.weight") == 0 || 
                             strcmp(tname, "output.weight") == 0)) {
                         continue;