diff --git a/paddlespeech/s2t/modules/attention.py b/paddlespeech/s2t/modules/attention.py
index 2d236743a6dabaa552f496ca2ed8bfa68b37e838..c02de15e801b19a40b8dd2a45609a3b1dd18c554 100644
--- a/paddlespeech/s2t/modules/attention.py
+++ b/paddlespeech/s2t/modules/attention.py
@@ -188,8 +188,9 @@ class MultiHeadedAttention(nn.Layer):
         #   non-trivial to calculate `next_cache_start` here.
         new_cache = paddle.concat((k, v), axis=-1)
 
-        scores = paddle.matmul(q,
-                               k.transpose([0, 1, 3, 2])) / math.sqrt(self.d_k)
+        # scores = paddle.matmul(q,
+                            #    k.transpose([0, 1, 3, 2])) / math.sqrt(self.d_k)
+        scores = paddle.matmul(q, k, transpose_y=True) / math.sqrt(self.d_k)
         return self.forward_attention(v, scores, mask), new_cache
 
 
@@ -309,11 +310,13 @@ class RelPositionMultiHeadedAttention(MultiHeadedAttention):
         # first compute matrix a and matrix c
         # as described in https://arxiv.org/abs/1901.02860 Section 3.3
         # (batch, head, time1, time2)
-        matrix_ac = paddle.matmul(q_with_bias_u, k.transpose([0, 1, 3, 2]))
+        # matrix_ac = paddle.matmul(q_with_bias_u, k.transpose([0, 1, 3, 2]))
+        matrix_ac = paddle.matmul(q_with_bias_u, k, transpose_y=True)
 
         # compute matrix b and matrix d
         # (batch, head, time1, time2)
-        matrix_bd = paddle.matmul(q_with_bias_v, p.transpose([0, 1, 3, 2]))
+        # matrix_bd = paddle.matmul(q_with_bias_v, p.transpose([0, 1, 3, 2]))
+        matrix_bd = paddle.matmul(q_with_bias_v, p, transpose_y=True)
         # Remove rel_shift since it is useless in speech recognition,
         # and it requires special attention for streaming.
         # matrix_bd = self.rel_shift(matrix_bd)