diff --git a/controllers/llamaCPP.cc b/controllers/llamaCPP.cc
index 45a0e49ba..291ec2ad7 100644
--- a/controllers/llamaCPP.cc
+++ b/controllers/llamaCPP.cc
@@ -197,7 +197,7 @@ void llamaCPP::chatCompletion(
     std::string grammar_file = (*jsonBody).get("grammar_file", "").asString();
     std::ifstream file(grammar_file);
     if (!file) {
-        LOG_ERROR << "Grammar file not found";
+      LOG_ERROR << "Grammar file not found";
     } else {
       std::stringstream grammarBuf;
       grammarBuf << file.rdbuf();
@@ -439,6 +439,14 @@ bool llamaCPP::loadModelImpl(const Json::Value &jsonBody) {
       LOG_INFO << "MMPROJ FILE detected, multi-model enabled!";
       params.mmproj = jsonBody["mmproj"].asString();
     }
+    if (!jsonBody["grp_attn_n"].isNull()) {
+
+      params.grp_attn_n = jsonBody["grp_attn_n"].asInt();
+    }
+    if (!jsonBody["grp_attn_w"].isNull()) {
+
+      params.grp_attn_w = jsonBody["grp_attn_w"].asInt();
+    }
     params.model = jsonBody["llama_model_path"].asString();
     params.n_gpu_layers = jsonBody.get("ngl", 100).asInt();
     params.n_ctx = jsonBody.get("ctx_len", 2048).asInt();
diff --git a/llama.cpp b/llama.cpp
index 326b418b5..862f5e41a 160000
--- a/llama.cpp
+++ b/llama.cpp
@@ -1 +1 @@
-Subproject commit 326b418b59b6d48d854c4461a2303e8ac0a311e6
+Subproject commit 862f5e41ab1fdf12d6f59455aad3f5dd8258f805