diff --git a/paddlespeech/t2s/modules/diffnet.py b/paddlespeech/t2s/modules/diffnet.py
index 2f433ad68..deae4847f 100644
--- a/paddlespeech/t2s/modules/diffnet.py
+++ b/paddlespeech/t2s/modules/diffnet.py
@@ -120,7 +120,11 @@ class SinusoidalPosEmb(nn.Layer):
         self.dim = dim
 
     def forward(self, x: paddle.Tensor):
-        x = paddle.cast(x, 'float32')
+        # check if x is 0-dim tensor, if so, add a dimension
+        if x.ndim == 0:
+            x = paddle.cast(x.unsqueeze(0), 'float32')
+        else:
+            x = paddle.cast(x, 'float32')
         half_dim = self.dim // 2
         emb = math.log(10000) / (half_dim - 1)
         emb = paddle.exp(paddle.arange(half_dim) * -emb)
diff --git a/paddlespeech/t2s/modules/nets_utils.py b/paddlespeech/t2s/modules/nets_utils.py
index 0a66a1c88..a3c6947b8 100644
--- a/paddlespeech/t2s/modules/nets_utils.py
+++ b/paddlespeech/t2s/modules/nets_utils.py
@@ -181,11 +181,12 @@ def make_pad_mask(lengths, xs=None, length_dim=-1):
     if length_dim == 0:
         raise ValueError("length_dim cannot be 0: {}".format(length_dim))
 
-    # check if ilens is 0-dim tensor, if so, add a dimension
+    # check if lengths is 0-dim tensor, if so, add a dimension
     if lengths.ndim == 0:
-        lengths = lengths.unsqueeze(0)
+        bs = paddle.shape(lengths.unsqueeze(0))
+    else:
+        bs = paddle.shape(lengths)
 
-    bs = paddle.shape(lengths)
     if xs is None:
         maxlen = paddle.cast(lengths.max(), dtype=bs.dtype)
     else: