From 120851df53e8a6c35a8c57704df4e1e4298f5245 Mon Sep 17 00:00:00 2001
From: Concedo <39025047+LostRuins@users.noreply.github.com>
Date: Mon, 12 Jun 2023 21:57:31 +0800
Subject: [PATCH] prevent gpu offload if kquant is selected with clblast for
 now

---
 llama.cpp | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/llama.cpp b/llama.cpp
index e6cd7e89b..eff3a6bb5 100644
--- a/llama.cpp
+++ b/llama.cpp
@@ -1077,7 +1077,11 @@ static void llama_model_load_internal(
     #if defined(GGML_USE_CLBLAST)
     if (file_version == LLAMA_FILE_VERSION_GGJT_V3) {
         if (hparams.ftype >= LLAMA_FTYPE_MOSTLY_Q2_K && hparams.ftype <= LLAMA_FTYPE_MOSTLY_Q6_K) {
-            printf("\n===\nK-Quants are currently not supported with CLBlast!!!\nPlease select a q4_0, q4_0, q5_0 or q5_1 format instead!\n=====\n");
+            if(n_gpu_layers>0)
+            {
+                n_gpu_layers = 0;
+                printf("\n===\nCLBlast cannot offload layers for K-Quants!\nPlease select a q4_0, q4_0, q5_0 or q5_1 format instead!\n=====\n");
+            }
         }
     }
     #endif