diff --git a/paddlespeech/s2t/modules/attention.py b/paddlespeech/s2t/modules/attention.py
index b2184dbc7..43700ca1e 100644
--- a/paddlespeech/s2t/modules/attention.py
+++ b/paddlespeech/s2t/modules/attention.py
@@ -26,10 +26,7 @@ from paddlespeech.s2t.utils.log import Log
 
 logger = Log(__name__).getlog()
 
-__all__ = [
-    "MultiHeadedAttention", "RelPositionMultiHeadedAttention",
-    "RelPositionMultiHeadedAttention2"
-]
+__all__ = ["MultiHeadedAttention", "RelPositionMultiHeadedAttention"]
 
 # Relative Positional Encodings
 # https://www.jianshu.com/p/c0608efcc26f
@@ -203,7 +200,10 @@ class MultiHeadedAttention(nn.Layer):
 class RelPositionMultiHeadedAttention(MultiHeadedAttention):
     """Multi-Head Attention layer with relative position encoding."""
 
-    def __init__(self, n_head, n_feat, dropout_rate,
+    def __init__(self,
+                 n_head,
+                 n_feat,
+                 dropout_rate,
                  do_rel_shift=False,
                  adaptive_scale=False,
                  init_weights=False):
diff --git a/paddlespeech/s2t/modules/conformer_convolution.py b/paddlespeech/s2t/modules/conformer_convolution.py
index e4196e3d4..7a0c72f3b 100644
--- a/paddlespeech/s2t/modules/conformer_convolution.py
+++ b/paddlespeech/s2t/modules/conformer_convolution.py
@@ -66,7 +66,6 @@ class ConvolutionModule(nn.Layer):
                 [1, 1, channels], default_initializer=I.Constant(0.0))
             self.add_parameter('ada_bias', ada_bias)
 
-
         self.pointwise_conv1 = Conv1D(
             channels,
             2 * channels,