diff --git a/ggml/src/ggml-zdnn/ggml-zdnn-rewrite.cpp b/ggml/src/ggml-zdnn/ggml-zdnn-rewrite.cpp
index aa16c10639..ff099fffed 100644
--- a/ggml/src/ggml-zdnn/ggml-zdnn-rewrite.cpp
+++ b/ggml/src/ggml-zdnn/ggml-zdnn-rewrite.cpp
@@ -47,6 +47,11 @@ inline void ggml_zdnn_create_tensor(zdnn_tensor_desc  & pre_tfm_desc,
     ZDNN_CHECK(zdnn_init_ztensor_with_malloc(&pre_tfm_desc, &tfm_desc, &ztensor));
 }
 
+inline void ggml_zdnn_load_tensor(zdnn_ztensor & ztensor,
+                                          void * buffer) {
+    ZDNN_CHECK(zdnn_transform_ztensor(&ztensor, buffer));
+}
+
 static void ggml_zdnn_mul_mat_op(ggml_backend_zdnn_context * ctx, const ggml_tensor * src0, const ggml_tensor * src1, ggml_tensor * dst) {
     GGML_TENSOR_BINARY_OP_LOCALS;
 
@@ -98,7 +103,10 @@ static void ggml_zdnn_mul_mat_op(ggml_backend_zdnn_context * ctx, const ggml_ten
     ggml_zdnn_create_tensor(ptd_output,  td_output,  zt_output,  output,  output_dim,  ZDNN_2D);
 
     void * bias_data = (void *)calloc(ne0, ggml_element_size(output));
-    ZDNN_CHECK(zdnn_transform_ztensor(&zt_bias, bias_data));
+    ggml_zdnn_load_tensor(zt_weights, weights->data);
+    ggml_zdnn_load_tensor(zt_inputs,  inputs->data);
+    ggml_zdnn_load_tensor(zt_bias,    bias_data);
+    ggml_zdnn_load_tensor(zt_output,  output->data);
 
     ZDNN_CHECK(zdnn_matmul_transpose_op(&zt_inputs, &zt_weights, &zt_bias,
                                         false, true, MATMUL_OP_ADDITION, &zt_output));