diff --git a/mindspore/nn/layer/math.py b/mindspore/nn/layer/math.py
index 15a3100650..3a3024628c 100644
--- a/mindspore/nn/layer/math.py
+++ b/mindspore/nn/layer/math.py
@@ -565,7 +565,7 @@ class IGamma(Cell):
     Above :math:`Q(a, x)` is the upper regularized complete Gamma function.
 
     Supported Platforms:
-        ``Ascend``
+        ``Ascend`` ``GPU``
 
     Inputs:
         - **a** (Tensor) - The input tensor. With float32 data type. `a` should have
@@ -619,8 +619,8 @@ class IGamma(Cell):
         use_igammac = self.logicaland(self.greater(x, 1), self.greater(x, a))
         ax = a * self.log(x) - x - self.lgamma(a)
         para_shape = self.shape(ax)
-        broadcastto = P.BroadcastTo(para_shape)
         if para_shape != ():
+            broadcastto = P.BroadcastTo(para_shape)
             x = broadcastto(x)
             a = broadcastto(a)
         x_is_zero = self.equal(x, 0)
@@ -694,8 +694,8 @@ class LBeta(Cell):
         _check_input_dtype("y", y_dtype, x_dtype, self.cls_name)
         x_plus_y = x + y
         para_shape = self.shape(x_plus_y)
-        broadcastto = P.BroadcastTo(para_shape)
         if para_shape != ():
+            broadcastto = P.BroadcastTo(para_shape)
             x = broadcastto(x)
             y = broadcastto(y)
         comp_less = self.less(x, y)