PaddlePaddle
diff --git a/‎python/paddle/nn/functional/norm.py
Copy file name to clipboard
+2-3Lines changed: 2 additions & 3 deletions b/‎python/paddle/nn/functional/norm.py
Copy file name to clipboard
+2-3Lines changed: 2 additions & 3 deletions
diff --git a/‎test/legacy_test/test_cross_entropy_loss.py
Copy file name to clipboard
+95-95Lines changed: 95 additions & 95 deletions b/‎test/legacy_test/test_cross_entropy_loss.py
Copy file name to clipboard
+95-95Lines changed: 95 additions & 95 deletions
diff --git a/‎test/legacy_test/test_imperative_auto_prune.py
Copy file name to clipboard
+32-32Lines changed: 32 additions & 32 deletions b/‎test/legacy_test/test_imperative_auto_prune.py
Copy file name to clipboard
+32-32Lines changed: 32 additions & 32 deletions
diff --git a/‎test/legacy_test/test_imperative_numpy_bridge.py
Copy file name to clipboard
+9-25Lines changed: 9 additions & 25 deletions b/‎test/legacy_test/test_imperative_numpy_bridge.py
Copy file name to clipboard
+9-25Lines changed: 9 additions & 25 deletions
@@ -14,9 +14,8 @@
 
 import numbers
 
-# TODO: define normalization api
 import paddle
-from paddle import _C_ops, base, in_dynamic_mode
+from paddle import _C_ops, in_dynamic_mode
 from paddle.base.framework import (
     in_dygraph_mode,
     in_dynamic_or_pir_mode,
@@ -83,7 +82,7 @@ def normalize(x, p=2, axis=1, epsilon=1e-12, name=None):
     """
 
     if in_dygraph_mode():
-        eps = base.dygraph.base.to_variable([epsilon], dtype=x.dtype)
+        eps = paddle.to_tensor([epsilon], dtype=x.dtype)
         out = _C_ops.p_norm(x, float(p), axis, epsilon, True, False)
         return x / _C_ops.maximum(out, eps)
 
 
@@ -171,8 +171,8 @@ def test_auto_prune(self):
             case1 = AutoPruneLayer0(input_size=5)
             value1 = np.arange(25).reshape(5, 5).astype("float32")
             value2 = np.arange(25).reshape(5, 5).astype("float32")
-            v1 = base.dygraph.to_variable(value1)
-            v2 = base.dygraph.to_variable(value2)
+            v1 = paddle.to_tensor(value1)
+            v2 = paddle.to_tensor(value2)
             loss = case1(v1, v2)
             loss.backward()
             self.assertIsNotNone(case1.linear2.weight._grad_ivar())
@@ -183,8 +183,8 @@ def test_auto_prune2(self):
             case2 = AutoPruneLayer1(input_size=5)
             value1 = np.arange(25).reshape(5, 5).astype("float32")
             value2 = np.arange(25).reshape(5, 5).astype("float32")
-            v1 = base.dygraph.to_variable(value1)
-            v2 = base.dygraph.to_variable(value2)
+            v1 = paddle.to_tensor(value1)
+            v2 = paddle.to_tensor(value2)
             loss = case2(v1, v2)
 
             loss.backward()
@@ -197,8 +197,8 @@ def test_auto_prune3(self):
             case3 = AutoPruneLayer3(input_size=784)
             value1 = np.arange(784).reshape(1, 784).astype("float32")
             value2 = np.arange(1).reshape(1, 1).astype("int64")
-            v1 = base.dygraph.to_variable(value1)
-            v2 = base.dygraph.to_variable(value2)
+            v1 = paddle.to_tensor(value1)
+            v2 = paddle.to_tensor(value2)
             loss, part2 = case3(v1, v2, 1)
             part2.retain_grads()
             loss.backward()
@@ -210,8 +210,8 @@ def test_auto_prune4(self):
             case4 = AutoPruneLayer3(input_size=784)
             value1 = np.arange(784).reshape(1, 784).astype("float32")
             value2 = np.arange(1).reshape(1, 1).astype("int64")
-            v1 = base.dygraph.to_variable(value1)
-            v2 = base.dygraph.to_variable(value2)
+            v1 = paddle.to_tensor(value1)
+            v2 = paddle.to_tensor(value2)
             loss, part2 = case4(v1, v2, 1)
             part2.retain_grads()
             part2.backward()
@@ -223,8 +223,8 @@ def test_auto_prune5(self):
             case4 = AutoPruneLayer3(input_size=784)
             value1 = np.arange(784).reshape(1, 784).astype("float32")
             value2 = np.arange(1).reshape(1, 1).astype("int64")
-            v1 = base.dygraph.to_variable(value1)
-            v2 = base.dygraph.to_variable(value2)
+            v1 = paddle.to_tensor(value1)
+            v2 = paddle.to_tensor(value2)
             loss, part1, part2 = case4(v1, v2, 2)
             part2.retain_grads()
             part1.backward()
@@ -238,9 +238,9 @@ def test_auto_prune6(self):
             value2 = np.arange(10).reshape(2, 5).astype("float32")
             linear = paddle.nn.Linear(13, 5)
             linear2 = paddle.nn.Linear(3, 3)
-            a = base.dygraph.to_variable(value0)
-            b = base.dygraph.to_variable(value1)
-            c = base.dygraph.to_variable(value2)
+            a = paddle.to_tensor(value0)
+            b = paddle.to_tensor(value1)
+            c = paddle.to_tensor(value2)
             out1 = linear(a)
             out2 = linear2(b)
             out1.stop_gradient = True
@@ -256,9 +256,9 @@ def test_auto_prune7(self):
             value2 = np.arange(10).reshape(2, 5).astype("float32")
             linear = paddle.nn.Linear(13, 5)
             linear2 = paddle.nn.Linear(3, 3)
-            a = base.dygraph.to_variable(value0)
-            b = base.dygraph.to_variable(value1)
-            c = base.dygraph.to_variable(value2)
+            a = paddle.to_tensor(value0)
+            b = paddle.to_tensor(value1)
+            c = paddle.to_tensor(value2)
             out1 = linear(a)
             out2 = linear2(b)
             out1.stop_gradient = True
@@ -274,9 +274,9 @@ def test_auto_prune8(self):
             value2 = np.arange(10).reshape(2, 5).astype("float32")
             linear = paddle.nn.Linear(13, 5)
             linear2 = paddle.nn.Linear(5, 3)
-            a = base.dygraph.to_variable(value0)
-            b = base.dygraph.to_variable(value1)
-            c = base.dygraph.to_variable(value2)
+            a = paddle.to_tensor(value0)
+            b = paddle.to_tensor(value1)
+            c = paddle.to_tensor(value2)
             out1 = linear(a)
             linear_origin = linear.weight.numpy()
             out2 = linear2(out1)
@@ -302,9 +302,9 @@ def test_auto_prune9(self):
             value2 = np.arange(10).reshape(2, 5).astype("float32")
             linear = paddle.nn.Linear(13, 5)
             linear2 = paddle.nn.Linear(5, 3)
-            a = base.dygraph.to_variable(value0)
-            b = base.dygraph.to_variable(value1)
-            c = base.dygraph.to_variable(value2)
+            a = paddle.to_tensor(value0)
+            b = paddle.to_tensor(value1)
+            c = paddle.to_tensor(value2)
             out1 = linear(a)
             linear_origin = linear.weight.numpy()
             out2 = linear2(out1)
@@ -332,9 +332,9 @@ def test_auto_prune10(self):
             value2 = np.arange(10).reshape(2, 5).astype("float32")
             linear = paddle.nn.Linear(13, 5)
             linear2 = paddle.nn.Linear(3, 3)
-            a = base.dygraph.to_variable(value0)
-            b = base.dygraph.to_variable(value1)
-            c = base.dygraph.to_variable(value2)
+            a = paddle.to_tensor(value0)
+            b = paddle.to_tensor(value1)
+            c = paddle.to_tensor(value2)
             out1 = linear(a)
             out2 = linear2(b)
             out1.stop_gradient = True
@@ -362,8 +362,8 @@ def test_auto_prune_with_optimizer(self):
             optimizer = paddle.optimizer.Adam(
                 0.001, parameters=model.parameters(), grad_clip=grad_clip
             )
-            indices = base.dygraph.to_variable(indices)
-            embed = base.dygraph.to_variable(embed)
+            indices = paddle.to_tensor(indices)
+            embed = paddle.to_tensor(embed)
             dummy_loss = model(embed)
 
             loss = model.embed_linear0(indices)
@@ -382,8 +382,8 @@ def test_auto_prune_with_optimizer(self):
                 0.001, parameters=model.parameters(), grad_clip=grad_clip
             )
 
-            indices = base.dygraph.to_variable(indices)
-            emebd = base.dygraph.to_variable(embed)
+            indices = paddle.to_tensor(indices)
+            emebd = paddle.to_tensor(embed)
             dummy_loss = model(indices)
 
             loss = model.embed_linear0(indices)
@@ -399,8 +399,8 @@ def test_case2_prune_no_grad_branch(self):
         with base.dygraph.guard():
             value1 = np.arange(784).reshape(1, 784)
             value2 = np.arange(1).reshape(1, 1)
-            v1 = base.dygraph.to_variable(value1).astype("float32")
-            v2 = base.dygraph.to_variable(value2).astype("float32")
+            v1 = paddle.to_tensor(value1).astype("float32")
+            v2 = paddle.to_tensor(value2).astype("float32")
             case3 = AutoPruneLayer2(input_size=784)
             loss = case3(v1, v2)
             loss.backward()
@@ -411,7 +411,7 @@ def test_case3_prune_no_grad_branch2(self):
         with base.dygraph.guard():
             value1 = np.arange(1).reshape(1, 1)
             linear = paddle.nn.Linear(1, 1)
-            label = base.dygraph.to_variable(value1).astype("float32")
+            label = paddle.to_tensor(value1).astype("float32")
             label = linear(label)
             label = paddle.cast(label, dtype="float32")
             label = paddle.cast(label, dtype='int64')
 
@@ -13,39 +13,23 @@
 # limitations under the License.
 
 import unittest
-import warnings
 
 import numpy as np
 
-from paddle import base
+import paddle
 
 
 class TestImperativeNumpyBridge(unittest.TestCase):
     def test_tensor_from_numpy(self):
         data_np = np.array([[2, 3, 1]]).astype('float32')
-        with base.dygraph.guard(base.CPUPlace()):
-            with warnings.catch_warnings(record=True) as w:
-                warnings.simplefilter("always")
-                var = base.dygraph.to_variable(data_np, zero_copy=True)
-                assert (
-                    "Currently, zero_copy is not supported, and it will be discarded."
-                    in str(w[-1].message)
-                )
-            # Temporally diable zero_copy
-            # var = base.dygraph.to_variable(data_np, zero_copy=True)
-            # np.testing.assert_array_equal(var.numpy(), data_np)
-            # data_np[0][0] = 4
-            # self.assertEqual(data_np[0][0], 4)
-            # self.assertEqual(var[0][0].numpy()[0], 4)
-            # np.testing.assert_array_equal(var.numpy(), data_np)
-
-            var2 = base.dygraph.to_variable(data_np, zero_copy=False)
-            np.testing.assert_array_equal(var2.numpy(), data_np)
-            data_np[0][0] = -1
-            self.assertEqual(data_np[0][0], -1)
-            self.assertNotEqual(var2[0][0].numpy(), -1)
-
-            self.assertFalse(np.array_equal(var2.numpy(), data_np))
+        paddle.set_device('cpu')
+        var2 = paddle.to_tensor(data_np)
+        np.testing.assert_array_equal(var2.numpy(), data_np)
+        data_np[0][0] = -1
+        self.assertEqual(data_np[0][0], -1)
+        self.assertNotEqual(var2[0][0].numpy(), -1)
+
+        self.assertFalse(np.array_equal(var2.numpy(), data_np))
 
 
 if __name__ == '__main__':