Fix q_xxs using mul_mat_q (#7459)

ggerganov · May 27, 2024 · 95f84d5 · 95f84d5
1 parent 5487593
commit 95f84d5
Showing 1 changed file with 1 addition and 0 deletions.
diff --git a/ggml-sycl.cpp b/ggml-sycl.cpp
@@ -15263,6 +15263,7 @@ static void ggml_sycl_mul_mat(const ggml_tensor * src0, const ggml_tensor * src1
             }
         } else {
             bool use_mul_mat_q = min_compute_capability >= VER_4VEC && ggml_is_quantized(src0->type);
+            use_mul_mat_q = use_mul_mat_q && (src0->type != GGML_TYPE_IQ2_XXS);
 
             if (use_xmx && min_compute_capability >= VER_GEN9 && src1->ne[1] > XMX_MAX_BATCH_SIZE) {
                 use_mul_mat_q = false;