use paddle Tensor instead of numpy array in pd/test_auto_batch_size.p…

…y to coverage newly added code

source/tests/pd/test_auto_batch_size.py

@@ -2,6 +2,7 @@
 import unittest
 
 import numpy as np
+import paddle
 
 from deepmd.pd.utils.auto_batch_size import (
     AutoBatchSize,
@@ -10,28 +11,28 @@
 
 class TestAutoBatchSize(unittest.TestCase):
     def test_execute_all(self):
-        dd0 = np.zeros((10000, 2, 1, 3, 4))
-        dd1 = np.ones((10000, 2, 1, 3, 4))
+        dd0 = paddle.zeros((10000, 2, 1, 3, 4))
+        dd1 = paddle.ones((10000, 2, 1, 3, 4))
         auto_batch_size = AutoBatchSize(256, 2.0)
 
         def func(dd1):
-            return np.zeros_like(dd1), np.ones_like(dd1)
+            return paddle.zeros_like(dd1), paddle.ones_like(dd1)
 
         dd2 = auto_batch_size.execute_all(func, 10000, 2, dd1)
-        np.testing.assert_equal(dd0, dd2[0])
-        np.testing.assert_equal(dd1, dd2[1])
+        np.testing.assert_equal(dd0.numpy(), dd2[0].numpy())
+        np.testing.assert_equal(dd1.numpy(), dd2[1].numpy())
 
     def test_execute_all_dict(self):
-        dd0 = np.zeros((10000, 2, 1, 3, 4))
-        dd1 = np.ones((10000, 2, 1, 3, 4))
+        dd0 = paddle.zeros((10000, 2, 1, 3, 4))
+        dd1 = paddle.ones((10000, 2, 1, 3, 4))
         auto_batch_size = AutoBatchSize(256, 2.0)
 
         def func(dd1):
             return {
-                "foo": np.zeros_like(dd1),
-                "bar": np.ones_like(dd1),
+                "foo": paddle.zeros_like(dd1),
+                "bar": paddle.ones_like(dd1),
             }
 
         dd2 = auto_batch_size.execute_all(func, 10000, 2, dd1)
-        np.testing.assert_equal(dd0, dd2["foo"])
-        np.testing.assert_equal(dd1, dd2["bar"])
+        np.testing.assert_equal(dd0.numpy(), dd2["foo"].numpy())
+        np.testing.assert_equal(dd1.numpy(), dd2["bar"].numpy())