API(GRUUnit) error message enhancement (#23535)

* err msg enhance for GRUUnit

* add ut
test=develop

python/paddle/fluid/dygraph/nn.py

@@ -1911,6 +1911,10 @@ class GRUUnit(layers.Layer):
                 self.activation, 'gate_activation', self.gate_activation)
             return updated_hidden, reset_hidden_pre, gate
 
+        check_variable_and_dtype(input, 'input', ['float32', 'float64'],
+                                 'GRUUnit')
+        check_variable_and_dtype(hidden, 'hidden', ['float32', 'float64'],
+                                 'GRUUnit')
         inputs = {
             'Input': [input],
             'HiddenPrev': [hidden],
@@ -1918,10 +1922,6 @@ class GRUUnit(layers.Layer):
         }
         if self.bias is not None:
             inputs['Bias'] = [self.bias]
-        attrs = {
-            'activation': self.activation,
-            'gate_activation': self.gate_activation,
-        }
         gate = self._helper.create_variable_for_type_inference(self._dtype)
         reset_hidden_pre = self._helper.create_variable_for_type_inference(
             self._dtype)

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

@@ -17,9 +17,25 @@ from __future__ import print_function
 import math
 import unittest
 import numpy as np
+import paddle.fluid as fluid
 from op_test import OpTest
 
 
+class TestGRUUnitAPIError(unittest.TestCase):
+    def test_errors(self):
+        with fluid.program_guard(fluid.Program(), fluid.Program()):
+            D = 5
+            layer = fluid.dygraph.nn.GRUUnit(size=D * 3)
+            # the input must be Variable.
+            x0 = fluid.create_lod_tensor(
+                np.array([-1, 3, 5, 5]), [[1, 1, 1, 1]], fluid.CPUPlace())
+            self.assertRaises(TypeError, layer, x0)
+            # the input dtype must be float32 or float64
+            x = fluid.data(name='x', shape=[-1, D * 3], dtype='float16')
+            hidden = fluid.data(name='hidden', shape=[-1, D], dtype='float32')
+            self.assertRaises(TypeError, layer, x, hidden)
+
+
 class GRUActivationType(OpTest):
     identity = 0
     sigmoid = 1