From c9983a72d6ee1924153e5f983ea27983dd091100 Mon Sep 17 00:00:00 2001
From: Concedo <39025047+LostRuins@users.noreply.github.com>
Date: Wed, 25 Oct 2023 15:18:03 +0800
Subject: [PATCH] prevent lora with clblast

---
 llama.cpp              | 5 +++--
 otherarch/llama_v3.cpp | 5 +++--
 2 files changed, 6 insertions(+), 4 deletions(-)

diff --git a/llama.cpp b/llama.cpp
index 249028ddf..76b47ccdc 100644
--- a/llama.cpp
+++ b/llama.cpp
@@ -8854,11 +8854,12 @@ static int llama_apply_lora_from_file_internal(
             offload_func_t offload_func = llama_nop;
             offload_func_t offload_func_force_inplace = llama_nop;
 
-#ifdef GGML_USE_CUBLAS
+#if defined(GGML_USE_CUBLAS) || defined(GGML_USE_CLBLAST)
             if (dest_t->backend == GGML_BACKEND_GPU || dest_t->backend == GGML_BACKEND_GPU_SPLIT) {
                 if (dest_t->type != GGML_TYPE_F16) {
+                    printf("\nError: the simultaneous use of LoRAs and GPU acceleration is only supported for f16 models\n");
                     throw std::runtime_error(format(
-                        "%s: error: the simultaneous use of LoRAs and GPU acceleration is only supported for f16 models", __func__));
+                        "%s: error: lora failed", __func__));
                 }
                 offload_func = ggml_cuda_assign_buffers;
                 offload_func_force_inplace = ggml_cuda_assign_buffers_force_inplace;
diff --git a/otherarch/llama_v3.cpp b/otherarch/llama_v3.cpp
index 703350a80..7eb3485f7 100644
--- a/otherarch/llama_v3.cpp
+++ b/otherarch/llama_v3.cpp
@@ -3763,11 +3763,12 @@ int llama_v3_apply_lora_from_file_internal(const struct llama_v3_model & model,
             offload_func_t offload_func = llama_v3_nop;
             offload_func_t offload_func_force_inplace = llama_v3_nop;
 
-#ifdef GGML_USE_CUBLAS
+#if defined(GGML_USE_CUBLAS) || defined(GGML_USE_CLBLAST)
             if (dest_t->backend == GGML_BACKEND_GPU || dest_t->backend == GGML_BACKEND_GPU_SPLIT) {
                 if (dest_t->type != GGML_TYPE_F16) {
+                    printf("\nError: the simultaneous use of LoRAs and GPU acceleration is only supported for f16 models\n");
                     throw std::runtime_error(format_old(
-                        "%s: error: the simultaneous use of LoRAs and GPU acceleration is only supported for f16 models", __func__));
+                        "%s: error: lora failed", __func__));
                 }
                 offload_func = ggml_cuda_assign_buffers;
                 offload_func_force_inplace = ggml_cuda_assign_buffers_force_inplace;