fix the indexerror of conv2d_transpose (#50005)

Author: Liyulingyue

python/paddle/fluid/tests/unittests/test_conv2d_transpose_op.py

@@ -989,6 +989,17 @@ def error_groups():
 
         self.assertRaises(ValueError, error_groups)
 
+        def error_0_filter_number():
+            out = paddle.static.nn.conv2d_transpose(
+                input=data,
+                groups=1,
+                num_filters=0,
+                filter_size=3,
+                data_format='NCHW',
+            )
+
+        self.assertRaises(ValueError, error_0_filter_number)
+
 
 class TestConv2DTransposeRepr(unittest.TestCase):
     def test_case(self):

python/paddle/static/nn/common.py

@@ -1542,6 +1542,9 @@ def conv2d_transpose(
             "but received {}".format(len(input.shape))
         )
 
+    if num_filters == 0:
+        raise ValueError("num of filters should not be 0.")
+
     if data_format not in ['NCHW', 'NHWC']:
         raise ValueError(
             "Attr(data_format) of Op(paddle.static.nn.layers.conv2d_transpose) got wrong value: received "