From 1c8ef38e56f0b864c323efaeacc817e31e2b8ecd Mon Sep 17 00:00:00 2001
From: Zhou Wei <1183042833@qq.com>
Date: Fri, 21 Oct 2022 00:30:31 +0800
Subject: [PATCH] [Sparse] change paddle.incubate.sparse to paddle.sparse
 (#47152)

---
 python/paddle/__init__.py                     |   1 +
 .../fluid/dygraph/varbase_patch_methods.py    |   4 +-
 .../tests/unittests/test_sparse_addmm_op.py   |   2 +-
 .../tests/unittests/test_sparse_conv_op.py    |  65 +++++-----
 .../unittests/test_sparse_elementwise_op.py   |  10 +-
 .../test_sparse_fused_attention_op.py         |   4 +-
 .../unittests/test_sparse_is_same_shape.py    |   4 +-
 .../tests/unittests/test_sparse_matmul_op.py  |   6 +-
 .../tests/unittests/test_sparse_model.py      |  12 +-
 .../tests/unittests/test_sparse_mv_op.py      |   4 +-
 .../tests/unittests/test_sparse_norm_op.py    |  14 +--
 .../tests/unittests/test_sparse_pooling_op.py |   6 +-
 .../tests/unittests/test_sparse_reshape_op.py |   8 +-
 .../tests/unittests/test_sparse_softmax_op.py |   4 +-
 .../unittests/test_sparse_transpose_op.py     |   2 +-
 .../tests/unittests/test_sparse_unary_op.py   |  57 ++++-----
 .../tests/unittests/test_sparse_utils_op.py   | 119 +++++++++---------
 python/paddle/incubate/__init__.py            |   1 -
 .../paddle/{incubate => }/sparse/__init__.py  |  37 +++++-
 python/paddle/{incubate => }/sparse/binary.py |  31 ++---
 .../paddle/{incubate => }/sparse/creation.py  |   4 +-
 .../paddle/{incubate => }/sparse/multiary.py  |   9 +-
 .../{incubate => }/sparse/nn/__init__.py      |   0
 .../sparse/nn/functional/__init__.py          |   0
 .../sparse/nn/functional/activation.py        |   8 +-
 .../sparse/nn/functional/conv.py              |   8 +-
 .../sparse/nn/functional/pooling.py           |   2 +-
 .../sparse/nn/functional/transformer.py       |   5 +-
 .../sparse/nn/layer/activation.py             |   8 +-
 .../{incubate => }/sparse/nn/layer/conv.py    |   8 +-
 .../{incubate => }/sparse/nn/layer/norm.py    |   8 +-
 .../{incubate => }/sparse/nn/layer/pooling.py |   2 +-
 python/paddle/{incubate => }/sparse/unary.py  |  46 ++++---
 python/setup.py.in                            |  17 ++-
 tools/print_signatures.py                     |   3 +
 35 files changed, 273 insertions(+), 246 deletions(-)
 rename python/paddle/{incubate => }/sparse/__init__.py (76%)
 rename python/paddle/{incubate => }/sparse/binary.py (93%)
 rename python/paddle/{incubate => }/sparse/creation.py (98%)
 rename python/paddle/{incubate => }/sparse/multiary.py (90%)
 rename python/paddle/{incubate => }/sparse/nn/__init__.py (100%)
 rename python/paddle/{incubate => }/sparse/nn/functional/__init__.py (100%)
 rename python/paddle/{incubate => }/sparse/nn/functional/activation.py (95%)
 rename python/paddle/{incubate => }/sparse/nn/functional/conv.py (97%)
 rename python/paddle/{incubate => }/sparse/nn/functional/pooling.py (97%)
 rename python/paddle/{incubate => }/sparse/nn/functional/transformer.py (95%)
 rename python/paddle/{incubate => }/sparse/nn/layer/activation.py (96%)
 rename python/paddle/{incubate => }/sparse/nn/layer/conv.py (98%)
 rename python/paddle/{incubate => }/sparse/nn/layer/norm.py (97%)
 rename python/paddle/{incubate => }/sparse/nn/layer/pooling.py (98%)
 rename python/paddle/{incubate => }/sparse/unary.py (93%)

diff --git a/python/paddle/__init__.py b/python/paddle/__init__.py
index 7312a0d9f8..9671520edd 100755
--- a/python/paddle/__init__.py
+++ b/python/paddle/__init__.py
@@ -82,6 +82,7 @@ import paddle.static  # noqa: F401
 import paddle.vision  # noqa: F401
 import paddle.audio  # noqa: F401
 import paddle.geometric  # noqa: F401
+import paddle.sparse  # noqa: F401
 
 from .tensor.attribute import is_complex  # noqa: F401
 from .tensor.attribute import is_integer  # noqa: F401
diff --git a/python/paddle/fluid/dygraph/varbase_patch_methods.py b/python/paddle/fluid/dygraph/varbase_patch_methods.py
index 61c856c071..618bcd540f 100644
--- a/python/paddle/fluid/dygraph/varbase_patch_methods.py
+++ b/python/paddle/fluid/dygraph/varbase_patch_methods.py
@@ -908,7 +908,7 @@ def monkey_patch_varbase():
                     indices = [[0, 0, 1, 2, 2], [1, 3, 2, 0, 1]]
                     values = [1, 2, 3, 4, 5]
                     dense_shape = [3, 4]
-                    sparse_x = paddle.incubate.sparse.sparse_coo_tensor(paddle.to_tensor(indices, dtype='int32'), paddle.to_tensor(values, dtype='float32'), shape=dense_shape)
+                    sparse_x = paddle.sparse.sparse_coo_tensor(paddle.to_tensor(indices, dtype='int32'), paddle.to_tensor(values, dtype='float32'), shape=dense_shape)
                     print(sparse_x.values())
                     #[1, 2, 3, 4, 5]
         """
@@ -933,7 +933,7 @@ def monkey_patch_varbase():
                     indices = [[0, 0, 1, 2, 2], [1, 3, 2, 0, 1]]
                     values = [1, 2, 3, 4, 5]
                     dense_shape = [3, 4]
-                    sparse_x = paddle.incubate.sparse.sparse_coo_tensor(paddle.to_tensor(indices, dtype='int64'), paddle.to_tensor(values, dtype='float32'), shape=dense_shape)
+                    sparse_x = paddle.sparse.sparse_coo_tensor(paddle.to_tensor(indices, dtype='int64'), paddle.to_tensor(values, dtype='float32'), shape=dense_shape)
                     dense_x = sparse_x.to_dense()
                     #[[0., 1., 0., 2.],
                     # [0., 0., 3., 0.],
diff --git a/python/paddle/fluid/tests/unittests/test_sparse_addmm_op.py b/python/paddle/fluid/tests/unittests/test_sparse_addmm_op.py
index 2e610d28dc..d93dba5c77 100644
--- a/python/paddle/fluid/tests/unittests/test_sparse_addmm_op.py
+++ b/python/paddle/fluid/tests/unittests/test_sparse_addmm_op.py
@@ -62,7 +62,7 @@ class TestAddmm(unittest.TestCase):
         sp_x.stop_gradient = False
         sp_y = origin_y.detach()
         sp_y.stop_gradient = False
-        sp_out = paddle.incubate.sparse.addmm(sp_input, sp_x, sp_y, 3.0, 2.0)
+        sp_out = paddle.sparse.addmm(sp_input, sp_x, sp_y, 3.0, 2.0)
 
         np.testing.assert_allclose(sp_out.numpy(),
                                    dense_out.numpy(),
diff --git a/python/paddle/fluid/tests/unittests/test_sparse_conv_op.py b/python/paddle/fluid/tests/unittests/test_sparse_conv_op.py
index 37418f5fa0..dfd7d23f8a 100644
--- a/python/paddle/fluid/tests/unittests/test_sparse_conv_op.py
+++ b/python/paddle/fluid/tests/unittests/test_sparse_conv_op.py
@@ -17,7 +17,7 @@ import numpy as np
 import paddle
 from paddle.fluid import core
 from paddle.fluid.framework import _test_eager_guard
-import paddle.incubate.sparse as sparse
+import paddle.sparse as sparse
 
 
 class TestSparseConv(unittest.TestCase):
@@ -42,17 +42,17 @@ class TestSparseConv(unittest.TestCase):
             correct_out_values = [[5], [11]]
             sparse_input = core.eager.sparse_coo_tensor(indices, values,
                                                         dense_shape, False)
-            out = paddle.incubate.sparse.nn.functional.conv3d(
-                sparse_input,
-                dense_kernel,
-                bias=paddle.to_tensor(bias, dtype='float32'),
-                stride=strides,
-                padding=paddings,
-                dilation=dilations,
-                groups=1,
-                data_format="NDHWC")
+            out = paddle.sparse.nn.functional.conv3d(sparse_input,
+                                                     dense_kernel,
+                                                     bias=paddle.to_tensor(
+                                                         bias, dtype='float32'),
+                                                     stride=strides,
+                                                     padding=paddings,
+                                                     dilation=dilations,
+                                                     groups=1,
+                                                     data_format="NDHWC")
             out.backward(out)
-            out = paddle.incubate.sparse.coalesce(out)
+            out = paddle.sparse.coalesce(out)
             assert np.array_equal(correct_out_values, out.values().numpy())
 
     def test_subm_conv3d(self):
@@ -62,11 +62,14 @@ class TestSparseConv(unittest.TestCase):
             indices = paddle.to_tensor(indices, dtype='int32')
             values = paddle.to_tensor(values, dtype='float32')
             dense_shape = [1, 1, 3, 4, 1]
-            sparse_x = paddle.incubate.sparse.sparse_coo_tensor(
-                indices, values, dense_shape, stop_gradient=True)
+            sparse_x = paddle.sparse.sparse_coo_tensor(indices,
+                                                       values,
+                                                       dense_shape,
+                                                       stop_gradient=True)
             weight = paddle.randn((1, 3, 3, 1, 1), dtype='float32')
-            y = paddle.incubate.sparse.nn.functional.subm_conv3d(
-                sparse_x, weight, key='subm_conv')
+            y = paddle.sparse.nn.functional.subm_conv3d(sparse_x,
+                                                        weight,
+                                                        key='subm_conv')
             assert np.array_equal(sparse_x.indices().numpy(),
                                   y.indices().numpy())
 
@@ -80,17 +83,20 @@ class TestSparseConv(unittest.TestCase):
             values = paddle.to_tensor(values, dtype='float32')
             dense_shape = [1, 1, 3, 4, 1]
             correct_out_values = [[4], [10]]
-            sparse_input = paddle.incubate.sparse.sparse_coo_tensor(
+            sparse_input = paddle.sparse.sparse_coo_tensor(
                 indices, values, dense_shape, False)
 
-            sparse_conv3d = paddle.incubate.sparse.nn.Conv3D(
-                1, 1, (1, 3, 3), data_format='NDHWC')
+            sparse_conv3d = paddle.sparse.nn.Conv3D(1,
+                                                    1, (1, 3, 3),
+                                                    data_format='NDHWC')
             sparse_out = sparse_conv3d(sparse_input)
             #test errors
             with self.assertRaises(ValueError):
                 #Currently, only support data_format='NDHWC'
-                conv3d = paddle.incubate.sparse.nn.SubmConv3D(
-                    1, 1, (1, 3, 3), data_format='NCDHW', key='subm_conv')
+                conv3d = paddle.sparse.nn.SubmConv3D(1,
+                                                     1, (1, 3, 3),
+                                                     data_format='NCDHW',
+                                                     key='subm_conv')
 
     def test_SubmConv3D(self):
         with _test_eager_guard():
@@ -100,11 +106,13 @@ class TestSparseConv(unittest.TestCase):
             values = paddle.to_tensor(values, dtype='float32')
             dense_shape = [1, 1, 3, 4, 1]
             correct_out_values = [[4], [10]]
-            sparse_input = paddle.incubate.sparse.sparse_coo_tensor(
+            sparse_input = paddle.sparse.sparse_coo_tensor(
                 indices, values, dense_shape, False)
 
-            subm_conv3d = paddle.incubate.sparse.nn.SubmConv3D(
-                1, 1, (1, 3, 3), data_format='NDHWC', key='subm_conv')
+            subm_conv3d = paddle.sparse.nn.SubmConv3D(1,
+                                                      1, (1, 3, 3),
+                                                      data_format='NDHWC',
+                                                      key='subm_conv')
             # test extra_repr
             print(subm_conv3d.extra_repr())
 
@@ -115,8 +123,10 @@ class TestSparseConv(unittest.TestCase):
             #test errors
             with self.assertRaises(ValueError):
                 #Currently, only support data_format='NDHWC'
-                conv3d = paddle.incubate.sparse.nn.SubmConv3D(
-                    1, 1, (1, 3, 3), data_format='NCDHW', key='subm_conv')
+                conv3d = paddle.sparse.nn.SubmConv3D(1,
+                                                     1, (1, 3, 3),
+                                                     data_format='NCDHW',
+                                                     key='subm_conv')
 
     def test_Conv3D_bias(self):
         with _test_eager_guard():
@@ -126,10 +136,7 @@ class TestSparseConv(unittest.TestCase):
             sp_x = x.to_sparse_coo(4)
             conv3d = paddle.nn.Conv3D(3, 2, 3, data_format='NDHWC')
 
-            sp_conv3d = paddle.incubate.sparse.nn.Conv3D(3,
-                                                         2,
-                                                         3,
-                                                         data_format='NDHWC')
+            sp_conv3d = paddle.sparse.nn.Conv3D(3, 2, 3, data_format='NDHWC')
             sp_conv3d.weight.set_value(
                 paddle.to_tensor(conv3d.weight.numpy().transpose(2, 3, 4, 1,
                                                                  0)))
diff --git a/python/paddle/fluid/tests/unittests/test_sparse_elementwise_op.py b/python/paddle/fluid/tests/unittests/test_sparse_elementwise_op.py
index c74a68b092..172bdaee2a 100644
--- a/python/paddle/fluid/tests/unittests/test_sparse_elementwise_op.py
+++ b/python/paddle/fluid/tests/unittests/test_sparse_elementwise_op.py
@@ -17,20 +17,20 @@ from operator import __add__, __sub__, __mul__, __truediv__
 
 import numpy as np
 import paddle
-import paddle.incubate.sparse as sparse
+import paddle.sparse as sparse
 
 op_list = [__add__, __sub__, __mul__, __truediv__]
 
 
 def get_actual_res(x, y, op):
     if op == __add__:
-        res = paddle.incubate.sparse.add(x, y)
+        res = paddle.sparse.add(x, y)
     elif op == __sub__:
-        res = paddle.incubate.sparse.subtract(x, y)
+        res = paddle.sparse.subtract(x, y)
     elif op == __mul__:
-        res = paddle.incubate.sparse.multiply(x, y)
+        res = paddle.sparse.multiply(x, y)
     elif op == __truediv__:
-        res = paddle.incubate.sparse.divide(x, y)
+        res = paddle.sparse.divide(x, y)
     else:
         raise ValueError("unsupported op")
     return res
diff --git a/python/paddle/fluid/tests/unittests/test_sparse_fused_attention_op.py b/python/paddle/fluid/tests/unittests/test_sparse_fused_attention_op.py
index 996ba3a011..50e9218a27 100644
--- a/python/paddle/fluid/tests/unittests/test_sparse_fused_attention_op.py
+++ b/python/paddle/fluid/tests/unittests/test_sparse_fused_attention_op.py
@@ -92,7 +92,7 @@ class TestSparseAttentionAPI1(unittest.TestCase):
                 output = paddle.matmul(softmax, value)
                 output.backward()
 
-                output_sp = paddle.incubate.sparse.nn.functional.attention(
+                output_sp = paddle.sparse.nn.functional.attention(
                     query_sp, key_sp, value_sp, sp_mask, kp_mask, attn_mask)
                 output_sp.backward()
             else:
@@ -103,7 +103,7 @@ class TestSparseAttentionAPI1(unittest.TestCase):
                 output = paddle.matmul(softmax, value)
                 output.backward()
 
-                output_sp = paddle.incubate.sparse.nn.functional.attention(
+                output_sp = paddle.sparse.nn.functional.attention(
                     query_sp, key_sp, value_sp, sp_mask)
                 output_sp.backward()
 
diff --git a/python/paddle/fluid/tests/unittests/test_sparse_is_same_shape.py b/python/paddle/fluid/tests/unittests/test_sparse_is_same_shape.py
index 8723d8da85..8715b55d52 100644
--- a/python/paddle/fluid/tests/unittests/test_sparse_is_same_shape.py
+++ b/python/paddle/fluid/tests/unittests/test_sparse_is_same_shape.py
@@ -15,12 +15,12 @@
 import unittest
 
 import paddle
-from paddle.incubate.sparse.binary import is_same_shape
+from paddle.sparse.binary import is_same_shape
 
 
 class TestSparseIsSameShapeAPI(unittest.TestCase):
     """
-    test paddle.incubate.sparse.is_same_shape
+    test paddle.sparse.is_same_shape
     """
 
     def setUp(self):
diff --git a/python/paddle/fluid/tests/unittests/test_sparse_matmul_op.py b/python/paddle/fluid/tests/unittests/test_sparse_matmul_op.py
index 7066951e5e..34a270b0fe 100644
--- a/python/paddle/fluid/tests/unittests/test_sparse_matmul_op.py
+++ b/python/paddle/fluid/tests/unittests/test_sparse_matmul_op.py
@@ -57,7 +57,7 @@ class TestMatmul(unittest.TestCase):
         sp_x.stop_gradient = False
         sp_y = origin_y.detach()
         sp_y.stop_gradient = False
-        sp_out = paddle.incubate.sparse.matmul(sp_x, sp_y)
+        sp_out = paddle.sparse.matmul(sp_x, sp_y)
 
         np.testing.assert_allclose(sp_out.numpy(),
                                    dense_out.numpy(),
@@ -106,7 +106,7 @@ class TestMaskedMatmul(unittest.TestCase):
         x = paddle.to_tensor(np_x, stop_gradient=False)
         y = paddle.to_tensor(np_y, stop_gradient=False)
         mask = paddle.to_tensor(np.ones([10, 6]) * np_mask).to_sparse_csr()
-        out = paddle.incubate.sparse.masked_matmul(x, y, mask)
+        out = paddle.sparse.masked_matmul(x, y, mask)
 
         np.testing.assert_allclose(np_out.indptr,
                                    out.crows().numpy(),
@@ -144,7 +144,7 @@ class TestMaskedMatmul(unittest.TestCase):
         sp_x.stop_gradient = False
         sp_y = origin_y.detach()
         sp_y.stop_gradient = False
-        sp_out = paddle.incubate.sparse.matmul(sp_x, sp_y)
+        sp_out = paddle.sparse.matmul(sp_x, sp_y)
         sp_out.backward()
 
         np.testing.assert_allclose(sp_out.numpy(),
diff --git a/python/paddle/fluid/tests/unittests/test_sparse_model.py b/python/paddle/fluid/tests/unittests/test_sparse_model.py
index c070614fc7..aa3d97919c 100644
--- a/python/paddle/fluid/tests/unittests/test_sparse_model.py
+++ b/python/paddle/fluid/tests/unittests/test_sparse_model.py
@@ -15,8 +15,7 @@
 import unittest
 import numpy as np
 import paddle
-from paddle.incubate import sparse
-from paddle.incubate.sparse import nn
+from paddle.sparse import nn
 from paddle.fluid.framework import _test_eager_guard
 
 
@@ -26,10 +25,10 @@ class TestGradientAdd(unittest.TestCase):
         indentity = sp_x
         out = nn.functional.relu(sp_x)
         values = out.values() + indentity.values()
-        out = sparse.sparse_coo_tensor(out.indices(),
-                                       values,
-                                       shape=out.shape,
-                                       stop_gradient=out.stop_gradient)
+        out = paddle.sparse.sparse_coo_tensor(out.indices(),
+                                              values,
+                                              shape=out.shape,
+                                              stop_gradient=out.stop_gradient)
         return out
 
     def dense(self, x):
@@ -55,7 +54,6 @@ class TestGradientAdd(unittest.TestCase):
             sparse_loss.backward(retain_graph=True)
 
             assert np.allclose(dense_out.numpy(), sparse_out.to_dense().numpy())
-            assert np.allclose(loss.numpy(), loss.numpy())
             assert np.allclose(x.grad.numpy(), sparse_x.grad.to_dense().numpy())
 
             loss.backward()
diff --git a/python/paddle/fluid/tests/unittests/test_sparse_mv_op.py b/python/paddle/fluid/tests/unittests/test_sparse_mv_op.py
index 277f78c670..e532063ad8 100644
--- a/python/paddle/fluid/tests/unittests/test_sparse_mv_op.py
+++ b/python/paddle/fluid/tests/unittests/test_sparse_mv_op.py
@@ -59,7 +59,7 @@ class TestCsrMv(unittest.TestCase):
             sp_x.stop_gradient = False
             sp_vec = origin_vec.detach()
             sp_vec.stop_gradient = False
-            sp_out = paddle.incubate.sparse.mv(sp_x, sp_vec)
+            sp_out = paddle.sparse.mv(sp_x, sp_vec)
             sp_out.backward()
 
             np.testing.assert_allclose(sp_out.numpy(),
@@ -97,7 +97,7 @@ class TestCooMv(unittest.TestCase):
             sp_x.stop_gradient = False
             sp_vec = origin_vec.detach()
             sp_vec.stop_gradient = False
-            sp_out = paddle.incubate.sparse.mv(sp_x, sp_vec)
+            sp_out = paddle.sparse.mv(sp_x, sp_vec)
             sp_out.backward()
 
             np.testing.assert_allclose(sp_out.numpy(),
diff --git a/python/paddle/fluid/tests/unittests/test_sparse_norm_op.py b/python/paddle/fluid/tests/unittests/test_sparse_norm_op.py
index 0959c1a1e4..503f171bdc 100644
--- a/python/paddle/fluid/tests/unittests/test_sparse_norm_op.py
+++ b/python/paddle/fluid/tests/unittests/test_sparse_norm_op.py
@@ -15,8 +15,8 @@
 import unittest
 import numpy as np
 import paddle
-from paddle.incubate.sparse import nn
-import paddle.incubate.sparse as sparse
+from paddle.sparse import nn
+import paddle.sparse as sparse
 import paddle.fluid as fluid
 import copy
 
@@ -40,7 +40,7 @@ class TestSparseBatchNorm(unittest.TestCase):
         dense_x2 = copy.deepcopy(dense_x)
         dense_x2.stop_gradient = False
         sparse_x = dense_x2.to_sparse_coo(sparse_dim)
-        sparse_batch_norm = paddle.incubate.sparse.nn.BatchNorm(channels)
+        sparse_batch_norm = paddle.sparse.nn.BatchNorm(channels)
         # set same params
         sparse_batch_norm._mean.set_value(batch_norm._mean)
         sparse_batch_norm._variance.set_value(batch_norm._variance)
@@ -66,8 +66,8 @@ class TestSparseBatchNorm(unittest.TestCase):
             shape = [2, 3, 6, 6, 3]
             x = paddle.randn(shape)
             sparse_x = x.to_sparse_coo(4)
-            sparse_batch_norm = paddle.incubate.sparse.nn.BatchNorm(
-                3, data_format='NCDHW')
+            sparse_batch_norm = paddle.sparse.nn.BatchNorm(3,
+                                                           data_format='NCDHW')
             sparse_batch_norm(sparse_x)
 
     def test2(self):
@@ -76,7 +76,7 @@ class TestSparseBatchNorm(unittest.TestCase):
         x_data = paddle.randn((1, 6, 6, 6, channels)).astype('float32')
         dense_x = paddle.to_tensor(x_data)
         sparse_x = dense_x.to_sparse_coo(4)
-        batch_norm = paddle.incubate.sparse.nn.BatchNorm(channels)
+        batch_norm = paddle.sparse.nn.BatchNorm(channels)
         batch_norm_out = batch_norm(sparse_x)
         dense_bn = paddle.nn.BatchNorm1D(channels)
         dense_x = dense_x.reshape((-1, dense_x.shape[-1]))
@@ -132,7 +132,7 @@ class TestStatic(unittest.TestCase):
         dense_shape = [1, 1, 3, 4, channels]
         sp_x = sparse.sparse_coo_tensor(indices, values, dense_shape)
 
-        sparse_batch_norm = paddle.incubate.sparse.nn.BatchNorm(channels)
+        sparse_batch_norm = paddle.sparse.nn.BatchNorm(channels)
         sp_y = sparse_batch_norm(sp_x)
         out = sp_y.to_dense()
 
diff --git a/python/paddle/fluid/tests/unittests/test_sparse_pooling_op.py b/python/paddle/fluid/tests/unittests/test_sparse_pooling_op.py
index 5a244dce26..998b5be9bb 100644
--- a/python/paddle/fluid/tests/unittests/test_sparse_pooling_op.py
+++ b/python/paddle/fluid/tests/unittests/test_sparse_pooling_op.py
@@ -45,7 +45,7 @@ class TestMaxPool3DFunc(unittest.TestCase):
             self.setUp()
             self.dense_x.stop_gradient = False
             sparse_x = self.dense_x.to_sparse_coo(4)
-            sparse_out = paddle.incubate.sparse.nn.functional.max_pool3d(
+            sparse_out = paddle.sparse.nn.functional.max_pool3d(
                 sparse_x,
                 self.kernel_sizes,
                 stride=self.strides,
@@ -106,8 +106,8 @@ class TestMaxPool3DAPI(unittest.TestCase):
         with _test_eager_guard():
             dense_x = paddle.randn((2, 3, 6, 6, 3))
             sparse_x = dense_x.to_sparse_coo(4)
-            max_pool3d = paddle.incubate.sparse.nn.MaxPool3D(
-                kernel_size=3, data_format='NDHWC')
+            max_pool3d = paddle.sparse.nn.MaxPool3D(kernel_size=3,
+                                                    data_format='NDHWC')
             out = max_pool3d(sparse_x)
             out = out.to_dense()
 
diff --git a/python/paddle/fluid/tests/unittests/test_sparse_reshape_op.py b/python/paddle/fluid/tests/unittests/test_sparse_reshape_op.py
index e9ef737f77..01619f76dc 100644
--- a/python/paddle/fluid/tests/unittests/test_sparse_reshape_op.py
+++ b/python/paddle/fluid/tests/unittests/test_sparse_reshape_op.py
@@ -19,7 +19,7 @@ import unittest
 
 class TestReshape(unittest.TestCase):
     """
-    Test the API paddle.incubate.sparse.reshape on some sparse tensors.
+    Test the API paddle.sparse.reshape on some sparse tensors.
     x: sparse, out: sparse
     """
 
@@ -31,7 +31,7 @@ class TestReshape(unittest.TestCase):
         Transform a sparse tensor with shape "x_shape" to
         a sparse tensor with shape "new_shape".
         Compare the output of paddle.reshape and the output of
-        paddle.incubate.sparse.reshape.
+        paddle.sparse.reshape.
         """
         mask = np.random.randint(0, 2, x_shape)
         np_x = np.random.randint(-100, 100, x_shape) * mask
@@ -49,7 +49,7 @@ class TestReshape(unittest.TestCase):
             sp_x = paddle.to_tensor(np_x,
                                     place=paddle.CPUPlace()).to_sparse_csr()
         sp_x.stop_gradient = False
-        sp_out = paddle.incubate.sparse.reshape(sp_x, new_shape)
+        sp_out = paddle.sparse.reshape(sp_x, new_shape)
 
         np.testing.assert_allclose(sp_out.to_dense().numpy(),
                                    dense_out.numpy(),
@@ -75,7 +75,7 @@ class TestReshape(unittest.TestCase):
                 sp_x = paddle.to_tensor(
                     np_x, place=paddle.CUDAPlace(0)).to_sparse_csr()
             sp_x.stop_gradient = False
-            sp_out = paddle.incubate.sparse.reshape(sp_x, new_shape)
+            sp_out = paddle.sparse.reshape(sp_x, new_shape)
 
             np.testing.assert_allclose(sp_out.to_dense().numpy(),
                                        dense_out.numpy(),
diff --git a/python/paddle/fluid/tests/unittests/test_sparse_softmax_op.py b/python/paddle/fluid/tests/unittests/test_sparse_softmax_op.py
index ede3375433..251f6cce1c 100644
--- a/python/paddle/fluid/tests/unittests/test_sparse_softmax_op.py
+++ b/python/paddle/fluid/tests/unittests/test_sparse_softmax_op.py
@@ -44,7 +44,7 @@ class TestCsrSoftmax(unittest.TestCase):
                 np_out = np.concatenate([np_out, x_exp / x_exp_sum])
 
             csr = paddle.to_tensor(np_x, stop_gradient=False).to_sparse_csr()
-            m = paddle.incubate.sparse.nn.Softmax()
+            m = paddle.sparse.nn.Softmax()
             out = m(csr)
             np.testing.assert_allclose(out.crows().numpy(),
                                        np_csr.indptr,
@@ -101,7 +101,7 @@ class TestCsrSoftmax(unittest.TestCase):
                     np_out = np.concatenate([np_out, x_exp / x_exp_sum])
 
             csr = paddle.to_tensor(np_x, stop_gradient=False).to_sparse_csr()
-            m = paddle.incubate.sparse.nn.Softmax()
+            m = paddle.sparse.nn.Softmax()
             out = m(csr)
             np.testing.assert_allclose(out.values().numpy(), np_out, rtol=1e-05)
 
diff --git a/python/paddle/fluid/tests/unittests/test_sparse_transpose_op.py b/python/paddle/fluid/tests/unittests/test_sparse_transpose_op.py
index 58bcbdc8c0..12d1fd9a8b 100644
--- a/python/paddle/fluid/tests/unittests/test_sparse_transpose_op.py
+++ b/python/paddle/fluid/tests/unittests/test_sparse_transpose_op.py
@@ -35,7 +35,7 @@ class TestTranspose(unittest.TestCase):
             else:
                 sp_x = origin_x.detach().to_sparse_csr()
             sp_x.stop_gradient = False
-            sp_out = paddle.incubate.sparse.transpose(sp_x, dims)
+            sp_out = paddle.sparse.transpose(sp_x, dims)
 
             np.testing.assert_allclose(sp_out.to_dense().numpy(),
                                        dense_out.numpy(),
diff --git a/python/paddle/fluid/tests/unittests/test_sparse_unary_op.py b/python/paddle/fluid/tests/unittests/test_sparse_unary_op.py
index 7abbaa84ad..fc39ab973c 100644
--- a/python/paddle/fluid/tests/unittests/test_sparse_unary_op.py
+++ b/python/paddle/fluid/tests/unittests/test_sparse_unary_op.py
@@ -90,84 +90,79 @@ class TestSparseUnary(unittest.TestCase):
         self.check_result(dense_func, sparse_func, 'csr', attr1, attr2)
 
     def test_sparse_sin(self):
-        self.compare_with_dense(paddle.sin, paddle.incubate.sparse.sin)
+        self.compare_with_dense(paddle.sin, paddle.sparse.sin)
 
     def test_sparse_tan(self):
-        self.compare_with_dense(paddle.tan, paddle.incubate.sparse.tan)
+        self.compare_with_dense(paddle.tan, paddle.sparse.tan)
 
     def test_sparse_asin(self):
-        self.compare_with_dense(paddle.asin, paddle.incubate.sparse.asin)
+        self.compare_with_dense(paddle.asin, paddle.sparse.asin)
 
     def test_sparse_atan(self):
-        self.compare_with_dense(paddle.atan, paddle.incubate.sparse.atan)
+        self.compare_with_dense(paddle.atan, paddle.sparse.atan)
 
     def test_sparse_sinh(self):
-        self.compare_with_dense(paddle.sinh, paddle.incubate.sparse.sinh)
+        self.compare_with_dense(paddle.sinh, paddle.sparse.sinh)
 
     def test_sparse_tanh(self):
-        self.compare_with_dense(paddle.tanh, paddle.incubate.sparse.tanh)
+        self.compare_with_dense(paddle.tanh, paddle.sparse.tanh)
 
     def test_sparse_asinh(self):
-        self.compare_with_dense(paddle.asinh, paddle.incubate.sparse.asinh)
+        self.compare_with_dense(paddle.asinh, paddle.sparse.asinh)
 
     def test_sparse_atanh(self):
-        self.compare_with_dense(paddle.atanh, paddle.incubate.sparse.atanh)
+        self.compare_with_dense(paddle.atanh, paddle.sparse.atanh)
 
     def test_sparse_sqrt(self):
-        self.compare_with_dense(paddle.sqrt, paddle.incubate.sparse.sqrt)
+        self.compare_with_dense(paddle.sqrt, paddle.sparse.sqrt)
 
     def test_sparse_square(self):
-        self.compare_with_dense(paddle.square, paddle.incubate.sparse.square)
+        self.compare_with_dense(paddle.square, paddle.sparse.square)
 
     def test_sparse_log1p(self):
-        self.compare_with_dense(paddle.log1p, paddle.incubate.sparse.log1p)
+        self.compare_with_dense(paddle.log1p, paddle.sparse.log1p)
 
     def test_sparse_relu(self):
-        self.compare_with_dense(paddle.nn.ReLU(),
-                                paddle.incubate.sparse.nn.ReLU())
+        self.compare_with_dense(paddle.nn.ReLU(), paddle.sparse.nn.ReLU())
 
     def test_sparse_relu6(self):
-        self.compare_with_dense(paddle.nn.ReLU6(),
-                                paddle.incubate.sparse.nn.ReLU6())
+        self.compare_with_dense(paddle.nn.ReLU6(), paddle.sparse.nn.ReLU6())
 
     def test_sparse_leaky_relu(self):
         self.compare_with_dense(paddle.nn.LeakyReLU(0.1),
-                                paddle.incubate.sparse.nn.LeakyReLU(0.1))
+                                paddle.sparse.nn.LeakyReLU(0.1))
 
     def test_sparse_abs(self):
-        self.compare_with_dense(paddle.abs, paddle.incubate.sparse.abs)
+        self.compare_with_dense(paddle.abs, paddle.sparse.abs)
 
     def test_sparse_expm1(self):
-        self.compare_with_dense(paddle.expm1, paddle.incubate.sparse.expm1)
+        self.compare_with_dense(paddle.expm1, paddle.sparse.expm1)
 
     def test_sparse_deg2rad(self):
-        self.compare_with_dense(paddle.deg2rad, paddle.incubate.sparse.deg2rad)
+        self.compare_with_dense(paddle.deg2rad, paddle.sparse.deg2rad)
 
     def test_sparse_rad2deg(self):
-        self.compare_with_dense(paddle.rad2deg, paddle.incubate.sparse.rad2deg)
+        self.compare_with_dense(paddle.rad2deg, paddle.sparse.rad2deg)
 
     def test_sparse_neg(self):
-        self.compare_with_dense(paddle.neg, paddle.incubate.sparse.neg)
+        self.compare_with_dense(paddle.neg, paddle.sparse.neg)
 
     def test_sparse_pow(self):
-        self.compare_with_dense_one_attr(paddle.pow, paddle.incubate.sparse.pow,
-                                         3)
+        self.compare_with_dense_one_attr(paddle.pow, paddle.sparse.pow, 3)
 
     def test_sparse_mul_scalar(self):
         self.compare_with_dense_one_attr(paddle.Tensor.__mul__,
-                                         paddle.incubate.sparse.multiply, 3)
+                                         paddle.sparse.multiply, 3)
 
     def test_sparse_div_scalar(self):
         self.compare_with_dense_one_attr(paddle.Tensor.__div__,
-                                         paddle.incubate.sparse.divide, 2)
+                                         paddle.sparse.divide, 2)
 
     def test_sparse_cast(self):
-        self.compare_with_dense_two_attr(paddle.cast,
-                                         paddle.incubate.sparse.cast, 'int32',
-                                         'float32')
-        self.compare_with_dense_two_attr(paddle.cast,
-                                         paddle.incubate.sparse.cast, 'int32',
-                                         'float64')
+        self.compare_with_dense_two_attr(paddle.cast, paddle.sparse.cast,
+                                         'int32', 'float32')
+        self.compare_with_dense_two_attr(paddle.cast, paddle.sparse.cast,
+                                         'int32', 'float64')
 
 
 if __name__ == "__main__":
diff --git a/python/paddle/fluid/tests/unittests/test_sparse_utils_op.py b/python/paddle/fluid/tests/unittests/test_sparse_utils_op.py
index 62a1da98d2..700c0795a0 100644
--- a/python/paddle/fluid/tests/unittests/test_sparse_utils_op.py
+++ b/python/paddle/fluid/tests/unittests/test_sparse_utils_op.py
@@ -31,10 +31,10 @@ class TestSparseCreate(unittest.TestCase):
             dense_shape = [3, 4]
             dense_indices = paddle.to_tensor(indices)
             dense_elements = paddle.to_tensor(values, dtype='float32')
-            coo = paddle.incubate.sparse.sparse_coo_tensor(dense_indices,
-                                                           dense_elements,
-                                                           dense_shape,
-                                                           stop_gradient=False)
+            coo = paddle.sparse.sparse_coo_tensor(dense_indices,
+                                                  dense_elements,
+                                                  dense_shape,
+                                                  stop_gradient=False)
             # test the to_string.py
             assert np.array_equal(indices, coo.indices().numpy())
             assert np.array_equal(values, coo.values().numpy())
@@ -44,8 +44,7 @@ class TestSparseCreate(unittest.TestCase):
             indices = [[0, 1, 2], [1, 2, 0]]
             values = [1.0, 2.0, 3.0]
             dense_shape = [3, 3]
-            coo = paddle.incubate.sparse.sparse_coo_tensor(
-                indices, values, dense_shape)
+            coo = paddle.sparse.sparse_coo_tensor(indices, values, dense_shape)
             assert np.array_equal(3, coo.nnz())
             assert np.array_equal(indices, coo.indices().numpy())
             assert np.array_equal(values, coo.values().numpy())
@@ -60,12 +59,11 @@ class TestSparseCreate(unittest.TestCase):
             dense_cols = paddle.to_tensor(cols)
             dense_elements = paddle.to_tensor(values, dtype='float32')
             stop_gradient = False
-            csr = paddle.incubate.sparse.sparse_csr_tensor(
-                dense_crows,
-                dense_cols,
-                dense_elements,
-                dense_shape,
-                stop_gradient=stop_gradient)
+            csr = paddle.sparse.sparse_csr_tensor(dense_crows,
+                                                  dense_cols,
+                                                  dense_elements,
+                                                  dense_shape,
+                                                  stop_gradient=stop_gradient)
 
     def test_create_csr_by_np(self):
         with _test_eager_guard():
@@ -73,8 +71,8 @@ class TestSparseCreate(unittest.TestCase):
             cols = [1, 3, 2, 0, 1]
             values = [1, 2, 3, 4, 5]
             dense_shape = [3, 4]
-            csr = paddle.incubate.sparse.sparse_csr_tensor(
-                crows, cols, values, dense_shape)
+            csr = paddle.sparse.sparse_csr_tensor(crows, cols, values,
+                                                  dense_shape)
             # test the to_string.py
             assert np.array_equal(5, csr.nnz())
             assert np.array_equal(crows, csr.crows().numpy())
@@ -87,10 +85,10 @@ class TestSparseCreate(unittest.TestCase):
             indices = [[0, 1], [0, 1]]
             values = [1.0, 2.0]
             dense_shape = [2, 2]
-            coo = paddle.incubate.sparse.sparse_coo_tensor(indices,
-                                                           values,
-                                                           dense_shape,
-                                                           place=place)
+            coo = paddle.sparse.sparse_coo_tensor(indices,
+                                                  values,
+                                                  dense_shape,
+                                                  place=place)
             assert coo.place.is_cpu_place()
             assert coo.values().place.is_cpu_place()
             assert coo.indices().place.is_cpu_place()
@@ -98,10 +96,10 @@ class TestSparseCreate(unittest.TestCase):
             crows = [0, 2, 3, 5]
             cols = [1, 3, 2, 0, 1]
             values = [1.0, 2.0, 3.0, 4.0, 5.0]
-            csr = paddle.incubate.sparse.sparse_csr_tensor(crows,
-                                                           cols,
-                                                           values, [3, 5],
-                                                           place=place)
+            csr = paddle.sparse.sparse_csr_tensor(crows,
+                                                  cols,
+                                                  values, [3, 5],
+                                                  place=place)
             assert csr.place.is_cpu_place()
             assert csr.crows().place.is_cpu_place()
             assert csr.cols().place.is_cpu_place()
@@ -114,19 +112,19 @@ class TestSparseCreate(unittest.TestCase):
             dense_shape = [2, 2]
             indices = paddle.to_tensor(indices, dtype='int32')
             values = paddle.to_tensor(values, dtype='float32')
-            coo = paddle.incubate.sparse.sparse_coo_tensor(indices,
-                                                           values,
-                                                           dense_shape,
-                                                           dtype='float64')
+            coo = paddle.sparse.sparse_coo_tensor(indices,
+                                                  values,
+                                                  dense_shape,
+                                                  dtype='float64')
             assert coo.dtype == paddle.float64
 
             crows = [0, 2, 3, 5]
             cols = [1, 3, 2, 0, 1]
             values = [1.0, 2.0, 3.0, 4.0, 5.0]
-            csr = paddle.incubate.sparse.sparse_csr_tensor(crows,
-                                                           cols,
-                                                           values, [3, 5],
-                                                           dtype='float16')
+            csr = paddle.sparse.sparse_csr_tensor(crows,
+                                                  cols,
+                                                  values, [3, 5],
+                                                  dtype='float16')
             assert csr.dtype == paddle.float16
 
     def test_create_coo_no_shape(self):
@@ -135,7 +133,7 @@ class TestSparseCreate(unittest.TestCase):
             values = [1.0, 2.0]
             indices = paddle.to_tensor(indices, dtype='int32')
             values = paddle.to_tensor(values, dtype='float32')
-            coo = paddle.incubate.sparse.sparse_coo_tensor(indices, values)
+            coo = paddle.sparse.sparse_coo_tensor(indices, values)
             assert [2, 2] == coo.shape
 
 
@@ -153,7 +151,7 @@ class TestSparseConvert(unittest.TestCase):
             #test to_sparse_coo_grad backward
             out_grad_indices = [[0, 1], [0, 1]]
             out_grad_values = [2.0, 3.0]
-            out_grad = paddle.incubate.sparse.sparse_coo_tensor(
+            out_grad = paddle.sparse.sparse_coo_tensor(
                 paddle.to_tensor(out_grad_indices),
                 paddle.to_tensor(out_grad_values),
                 shape=out.shape,
@@ -169,7 +167,7 @@ class TestSparseConvert(unittest.TestCase):
             values = [1.0, 2.0, 3.0, 4.0, 5.0]
             indices_dtypes = ['int32', 'int64']
             for indices_dtype in indices_dtypes:
-                sparse_x = paddle.incubate.sparse.sparse_coo_tensor(
+                sparse_x = paddle.sparse.sparse_coo_tensor(
                     paddle.to_tensor(indices, dtype=indices_dtype),
                     paddle.to_tensor(values),
                     shape=[3, 4],
@@ -185,7 +183,7 @@ class TestSparseConvert(unittest.TestCase):
                                       sparse_x.grad.values().numpy())
 
                 paddle.device.set_device("cpu")
-                sparse_x_cpu = paddle.incubate.sparse.sparse_coo_tensor(
+                sparse_x_cpu = paddle.sparse.sparse_coo_tensor(
                     paddle.to_tensor(indices, dtype=indices_dtype),
                     paddle.to_tensor(values),
                     shape=[3, 4],
@@ -216,7 +214,7 @@ class TestSparseConvert(unittest.TestCase):
         with _test_eager_guard():
             indices = [[0, 0, 1, 2, 2], [1, 3, 2, 0, 1]]
             values = [1.0, 2.0, 3.0, 4.0, 5.0]
-            sparse_x = paddle.incubate.sparse.sparse_coo_tensor(
+            sparse_x = paddle.sparse.sparse_coo_tensor(
                 paddle.to_tensor(indices),
                 paddle.to_tensor(values),
                 shape=[3, 4],
@@ -229,7 +227,7 @@ class TestSparseConvert(unittest.TestCase):
             indices = [[0, 0, 1, 2, 2], [1, 3, 2, 0, 1]]
             values = [[1.0, 1.0], [2.0, 2.0], [3.0, 3.0], [4.0, 4.0],
                       [5.0, 5.0]]
-            sparse_x = paddle.incubate.sparse.sparse_coo_tensor(
+            sparse_x = paddle.sparse.sparse_coo_tensor(
                 paddle.to_tensor(indices),
                 paddle.to_tensor(values),
                 shape=[3, 4, 2],
@@ -254,13 +252,13 @@ class TestSparseConvert(unittest.TestCase):
                     values = paddle.to_tensor(values,
                                               dtype='float32',
                                               stop_gradient=False)
-                    sparse_x = paddle.incubate.sparse.sparse_coo_tensor(
+                    sparse_x = paddle.sparse.sparse_coo_tensor(
                         indices, values, shape=[2, 2], stop_gradient=False)
                     grad_indices = [[0, 1], [1, 1]]
                     grad_values = [2, 3]
                     grad_indices = paddle.to_tensor(grad_indices, dtype='int32')
                     grad_values = paddle.to_tensor(grad_values, dtype='float32')
-                    sparse_out_grad = paddle.incubate.sparse.sparse_coo_tensor(
+                    sparse_out_grad = paddle.sparse.sparse_coo_tensor(
                         grad_indices, grad_values, shape=[2, 2])
                     sparse_x.backward(sparse_out_grad)
                     correct_values_grad = [0, 3]
@@ -272,11 +270,11 @@ class TestSparseConvert(unittest.TestCase):
                     values = paddle.to_tensor(values,
                                               dtype='float32',
                                               stop_gradient=False)
-                    sparse_x = paddle.incubate.sparse.sparse_coo_tensor(
+                    sparse_x = paddle.sparse.sparse_coo_tensor(
                         indices, values, shape=[2, 2, 2], stop_gradient=False)
                     grad_values = [[2, 2], [3, 3]]
                     grad_values = paddle.to_tensor(grad_values, dtype='float32')
-                    sparse_out_grad = paddle.incubate.sparse.sparse_coo_tensor(
+                    sparse_out_grad = paddle.sparse.sparse_coo_tensor(
                         grad_indices, grad_values, shape=[2, 2, 2])
                     sparse_x.backward(sparse_out_grad)
                     correct_values_grad = [[0, 0], [3, 3]]
@@ -294,9 +292,8 @@ class TestSparseConvert(unittest.TestCase):
                     values = [1.0, 2.0, 3.0]
                     indices = paddle.to_tensor(indices, dtype='int32')
                     values = paddle.to_tensor(values, dtype='float32')
-                    sparse_x = paddle.incubate.sparse.sparse_coo_tensor(
-                        indices, values)
-                    sparse_x = paddle.incubate.sparse.coalesce(sparse_x)
+                    sparse_x = paddle.sparse.sparse_coo_tensor(indices, values)
+                    sparse_x = paddle.sparse.coalesce(sparse_x)
                     indices_sorted = [[0, 1], [1, 0]]
                     values_sorted = [5.0, 1.0]
                     assert np.array_equal(indices_sorted,
@@ -307,9 +304,8 @@ class TestSparseConvert(unittest.TestCase):
                     # test the non-zero values is a vector
                     values = [[1.0, 1.0], [2.0, 2.0], [3.0, 3.0]]
                     values = paddle.to_tensor(values, dtype='float32')
-                    sparse_x = paddle.incubate.sparse.sparse_coo_tensor(
-                        indices, values)
-                    sparse_x = paddle.incubate.sparse.coalesce(sparse_x)
+                    sparse_x = paddle.sparse.sparse_coo_tensor(indices, values)
+                    sparse_x = paddle.sparse.coalesce(sparse_x)
                     values_sorted = [[5.0, 5.0], [1.0, 1.0]]
                     assert np.array_equal(indices_sorted,
                                           sparse_x.indices().numpy())
@@ -363,8 +359,9 @@ class TestCooError(unittest.TestCase):
                 values = [1, 2]
                 # 1. the shape too small
                 dense_shape = [2, 2]
-                sparse_x = paddle.incubate.sparse.sparse_coo_tensor(
-                    indices, values, shape=dense_shape)
+                sparse_x = paddle.sparse.sparse_coo_tensor(indices,
+                                                           values,
+                                                           shape=dense_shape)
 
     def test_same_nnz(self):
         with _test_eager_guard():
@@ -372,8 +369,7 @@ class TestCooError(unittest.TestCase):
                 # 2. test the nnz of indices must same as nnz of values
                 indices = [[1, 2], [1, 0]]
                 values = [1, 2, 3]
-                sparse_x = paddle.incubate.sparse.sparse_coo_tensor(
-                    indices, values)
+                sparse_x = paddle.sparse.sparse_coo_tensor(indices, values)
 
     def test_same_dimensions(self):
         with _test_eager_guard():
@@ -381,17 +377,16 @@ class TestCooError(unittest.TestCase):
                 indices = [[1, 2], [1, 0]]
                 values = [1, 2, 3]
                 shape = [2, 3, 4]
-                sparse_x = paddle.incubate.sparse.sparse_coo_tensor(indices,
-                                                                    values,
-                                                                    shape=shape)
+                sparse_x = paddle.sparse.sparse_coo_tensor(indices,
+                                                           values,
+                                                           shape=shape)
 
     def test_indices_dtype(self):
         with _test_eager_guard():
             with self.assertRaises(TypeError):
                 indices = [[1.0, 2.0], [0, 1]]
                 values = [1, 2]
-                sparse_x = paddle.incubate.sparse.sparse_coo_tensor(
-                    indices, values)
+                sparse_x = paddle.sparse.sparse_coo_tensor(indices, values)
 
 
 class TestCsrError(unittest.TestCase):
@@ -403,7 +398,7 @@ class TestCsrError(unittest.TestCase):
                 cols = [0, 1, 2]
                 values = [1, 2, 3]
                 shape = [3]
-                sparse_x = paddle.incubate.sparse.sparse_csr_tensor(
+                sparse_x = paddle.sparse.sparse_csr_tensor(
                     crows, cols, values, shape)
 
     def test_dimension2(self):
@@ -413,7 +408,7 @@ class TestCsrError(unittest.TestCase):
                 cols = [0, 1, 2]
                 values = [1, 2, 3]
                 shape = [3, 3, 3, 3]
-                sparse_x = paddle.incubate.sparse.sparse_csr_tensor(
+                sparse_x = paddle.sparse.sparse_csr_tensor(
                     crows, cols, values, shape)
 
     def test_same_shape1(self):
@@ -423,7 +418,7 @@ class TestCsrError(unittest.TestCase):
                 cols = [0, 1, 2, 3]
                 values = [1, 2, 3]
                 shape = [3, 4]
-                sparse_x = paddle.incubate.sparse.sparse_csr_tensor(
+                sparse_x = paddle.sparse.sparse_csr_tensor(
                     crows, cols, values, shape)
 
     def test_same_shape2(self):
@@ -433,7 +428,7 @@ class TestCsrError(unittest.TestCase):
                 cols = [0, 1, 2, 3]
                 values = [1, 2, 3, 4]
                 shape = [3, 4]
-                sparse_x = paddle.incubate.sparse.sparse_csr_tensor(
+                sparse_x = paddle.sparse.sparse_csr_tensor(
                     crows, cols, values, shape)
 
     def test_same_shape3(self):
@@ -443,7 +438,7 @@ class TestCsrError(unittest.TestCase):
                 cols = [0, 1, 2, 3, 0, 1, 2]
                 values = [1, 2, 3, 4, 0, 1, 2]
                 shape = [2, 3, 4]
-                sparse_x = paddle.incubate.sparse.sparse_csr_tensor(
+                sparse_x = paddle.sparse.sparse_csr_tensor(
                     crows, cols, values, shape)
 
     def test_crows_first_value(self):
@@ -453,7 +448,7 @@ class TestCsrError(unittest.TestCase):
                 cols = [0, 1, 2]
                 values = [1, 2, 3]
                 shape = [3, 4]
-                sparse_x = paddle.incubate.sparse.sparse_csr_tensor(
+                sparse_x = paddle.sparse.sparse_csr_tensor(
                     crows, cols, values, shape)
 
     def test_dtype(self):
@@ -463,7 +458,7 @@ class TestCsrError(unittest.TestCase):
                 cols = [0, 1, 2]
                 values = [1, 2, 3]
                 shape = [3]
-                sparse_x = paddle.incubate.sparse.sparse_csr_tensor(
+                sparse_x = paddle.sparse.sparse_csr_tensor(
                     crows, cols, values, shape)
 
 
diff --git a/python/paddle/incubate/__init__.py b/python/paddle/incubate/__init__.py
index 543b0b815c..68c7db0549 100644
--- a/python/paddle/incubate/__init__.py
+++ b/python/paddle/incubate/__init__.py
@@ -31,7 +31,6 @@ from .passes import fuse_resnet_unit_pass
 
 from . import autograd  #noqa: F401
 from . import autotune  #noqa: F401
-from . import sparse  #noqa: F401
 from . import nn  #noqa: F401
 from . import asp  #noqa: F401
 
diff --git a/python/paddle/incubate/sparse/__init__.py b/python/paddle/sparse/__init__.py
similarity index 76%
rename from python/paddle/incubate/sparse/__init__.py
rename to python/paddle/sparse/__init__.py
index 8b6866fa4d..9ca932ac46 100644
--- a/python/paddle/incubate/sparse/__init__.py
+++ b/python/paddle/sparse/__init__.py
@@ -51,9 +51,36 @@ from .multiary import addmm
 from . import nn
 
 __all__ = [
-    'sparse_coo_tensor', 'sparse_csr_tensor', 'sin', 'tan', 'asin', 'atan',
-    'sinh', 'tanh', 'asinh', 'atanh', 'sqrt', 'square', 'log1p', 'abs', 'pow',
-    'cast', 'neg', 'deg2rad', 'rad2deg', 'expm1', 'mv', 'matmul',
-    'masked_matmul', 'addmm', 'add', 'subtract', 'transpose', 'multiply',
-    'divide', 'coalesce', 'is_same_shape', 'reshape'
+    'sparse_coo_tensor',
+    'sparse_csr_tensor',
+    'sin',
+    'tan',
+    'asin',
+    'atan',
+    'sinh',
+    'tanh',
+    'asinh',
+    'atanh',
+    'sqrt',
+    'square',
+    'log1p',
+    'abs',
+    'pow',
+    'cast',
+    'neg',
+    'deg2rad',
+    'rad2deg',
+    'expm1',
+    'mv',
+    'matmul',
+    'masked_matmul',
+    'addmm',
+    'add',
+    'subtract',
+    'transpose',
+    'multiply',
+    'divide',
+    'coalesce',
+    'is_same_shape',
+    'reshape',
 ]
diff --git a/python/paddle/incubate/sparse/binary.py b/python/paddle/sparse/binary.py
similarity index 93%
rename from python/paddle/incubate/sparse/binary.py
rename to python/paddle/sparse/binary.py
index 656e0fd8dd..1e7fcfe763 100644
--- a/python/paddle/incubate/sparse/binary.py
+++ b/python/paddle/sparse/binary.py
@@ -64,19 +64,20 @@ def matmul(x, y, name=None):
 
         .. code-block:: python
 
+            # required: gpu
             import paddle
 
             # csr @ dense -> dense
             crows = [0, 1, 2, 3]
             cols = [1, 2, 0]
             values = [1., 2., 3.]
-            csr = paddle.incubate.sparse.sparse_csr_tensor(crows, cols, values, [3, 3])
+            csr = paddle.sparse.sparse_csr_tensor(crows, cols, values, [3, 3])
             # Tensor(shape=[3, 3], dtype=paddle.float32, place=Place(gpu:0), stop_gradient=True,
             #        crows=[0, 1, 2, 3],
             #        cols=[1, 2, 0],
             #        values=[1., 2., 3.])
             dense = paddle.ones([3, 2])
-            out = paddle.incubate.sparse.matmul(csr, dense)
+            out = paddle.sparse.matmul(csr, dense)
             # Tensor(shape=[3, 2], dtype=float32, place=Place(gpu:0), stop_gradient=True,
             #        [[1., 1.],
             #         [2., 2.],
@@ -85,13 +86,13 @@ def matmul(x, y, name=None):
             # coo @ dense -> dense
             indices = [[0, 1, 2], [1, 2, 0]]
             values = [1., 2., 3.]
-            coo = paddle.incubate.sparse.sparse_coo_tensor(indices, values, [3, 3])
+            coo = paddle.sparse.sparse_coo_tensor(indices, values, [3, 3])
             # Tensor(shape=[3, 3], dtype=paddle.float32, place=Place(gpu:0), stop_gradient=True,
             #        indices=[[0, 1, 2],
             #                 [1, 2, 0]],
             #        values=[1., 2., 3.])
             dense = paddle.ones([3, 2])
-            out = paddle.incubate.sparse.matmul(coo, dense)
+            out = paddle.sparse.matmul(coo, dense)
             # Tensor(shape=[3, 2], dtype=float32, place=Place(gpu:0), stop_gradient=True,
             #        [[1., 1.],
             #         [2., 2.],
@@ -133,6 +134,7 @@ def masked_matmul(x, y, mask, name=None):
 
         .. code-block:: python
 
+            # required: gpu
             import paddle
             paddle.seed(100)
 
@@ -141,7 +143,7 @@ def masked_matmul(x, y, mask, name=None):
             cols = [1, 3, 2, 0, 1]
             values = [1., 2., 3., 4., 5.]
             dense_shape = [3, 4]
-            mask = paddle.incubate.sparse.sparse_csr_tensor(crows, cols, values, dense_shape)
+            mask = paddle.sparse.sparse_csr_tensor(crows, cols, values, dense_shape)
             # Tensor(shape=[3, 4], dtype=paddle.float32, place=Place(gpu:0), stop_gradient=True,
             #       crows=[0, 2, 3, 5],
             #       cols=[1, 3, 2, 0, 1],
@@ -150,7 +152,7 @@ def masked_matmul(x, y, mask, name=None):
             x = paddle.rand([3, 5])
             y = paddle.rand([5, 4])
 
-            out = paddle.incubate.sparse.masked_matmul(x, y, mask)
+            out = paddle.sparse.masked_matmul(x, y, mask)
             # Tensor(shape=[3, 4], dtype=paddle.float32, place=Place(gpu:0), stop_gradient=True,
             #        crows=[0, 2, 3, 5],
             #        cols=[1, 3, 2, 0, 1],
@@ -191,6 +193,7 @@ def mv(x, vec, name=None):
 
         .. code-block:: python
 
+            # required: gpu
             import paddle
             paddle.seed(100)
 
@@ -199,14 +202,14 @@ def mv(x, vec, name=None):
             cols = [1, 3, 2, 0, 1]
             values = [1., 2., 3., 4., 5.]
             dense_shape = [3, 4]
-            csr = paddle.incubate.sparse.sparse_csr_tensor(crows, cols, values, dense_shape)
+            csr = paddle.sparse.sparse_csr_tensor(crows, cols, values, dense_shape)
             # Tensor(shape=[3, 4], dtype=paddle.float32, place=Place(gpu:0), stop_gradient=True,
             #        crows=[0, 2, 3, 5],
             #        cols=[1, 3, 2, 0, 1],
             #        values=[1., 2., 3., 4., 5.])
             vec = paddle.randn([4])
 
-            out = paddle.incubate.sparse.mv(csr, vec)
+            out = paddle.sparse.mv(csr, vec)
             # Tensor(shape=[3], dtype=float32, place=Place(gpu:0), stop_gradient=True,
             #        [-3.85499096, -2.42975140, -1.75087738])
 
@@ -245,7 +248,7 @@ def add(x, y, name=None):
             y = paddle.to_tensor([[0, 0, 0, -2], [0, 2, -3, 0], [2, 3, 4, 8]], 'float32')
             sparse_x = x.to_sparse_csr()
             sparse_y = y.to_sparse_csr()
-            sparse_z = paddle.incubate.sparse.add(sparse_x, sparse_y)
+            sparse_z = paddle.sparse.add(sparse_x, sparse_y)
             print(sparse_z.to_dense())
 
         # [[ 0., -1.,  0.,  0.],
@@ -302,7 +305,7 @@ def subtract(x, y, name=None):
             y = paddle.to_tensor([[0, 0, 0, -2], [0, 2, -3, 0], [2, 3, 4, 8]], 'float32')
             sparse_x = x.to_sparse_csr()
             sparse_y = y.to_sparse_csr()
-            sparse_z = paddle.incubate.sparse.subtract(sparse_x, sparse_y)
+            sparse_z = paddle.sparse.subtract(sparse_x, sparse_y)
             print(sparse_z.to_dense())
 
         # [[ 0., -1.,  0.,  4.],
@@ -347,7 +350,7 @@ def multiply(x, y, name=None):
             y = paddle.to_tensor([[0, 0, 0, -2], [0, 2, -3, 0], [2, 3, 4, 8]], 'float32')
             sparse_x = x.to_sparse_csr()
             sparse_y = y.to_sparse_csr()
-            sparse_z = paddle.incubate.sparse.multiply(sparse_x, sparse_y)
+            sparse_z = paddle.sparse.multiply(sparse_x, sparse_y)
             print(sparse_z.to_dense())
 
         # [[ 0.,  0.,  0., -4.],
@@ -395,7 +398,7 @@ def divide(x, y, name=None):
             y = paddle.to_tensor([[0, 0, 0, -2], [0, 2, -3, 0], [2, 3, 4, 8]], 'float32')
             sparse_x = x.to_sparse_csr()
             sparse_y = y.to_sparse_csr()
-            sparse_z = paddle.incubate.sparse.divide(sparse_x, sparse_y)
+            sparse_z = paddle.sparse.divide(sparse_x, sparse_y)
             print(sparse_z.to_dense())
 
         # [[ nan      , -inf.     ,  nan      , -1.       ],
@@ -438,9 +441,9 @@ def is_same_shape(x, y):
             y = y.to_sparse_csr()
             z = paddle.rand([2, 5])
 
-            paddle.incubate.sparse.is_same_shape(x, y)
+            paddle.sparse.is_same_shape(x, y)
             # True
-            paddle.incubate.sparse.is_same_shape(x, z)
+            paddle.sparse.is_same_shape(x, z)
             # False
 
     """
diff --git a/python/paddle/incubate/sparse/creation.py b/python/paddle/sparse/creation.py
similarity index 98%
rename from python/paddle/incubate/sparse/creation.py
rename to python/paddle/sparse/creation.py
index fe35b9cbd6..5ca050be08 100644
--- a/python/paddle/incubate/sparse/creation.py
+++ b/python/paddle/sparse/creation.py
@@ -107,7 +107,7 @@ def sparse_coo_tensor(indices,
             indices = [[0, 1, 2], [1, 2, 0]]
             values = [1.0, 2.0, 3.0]
             dense_shape = [3, 3]
-            coo = paddle.incubate.sparse.sparse_coo_tensor(indices, values, dense_shape)
+            coo = paddle.sparse.sparse_coo_tensor(indices, values, dense_shape)
             # print(coo)
             # Tensor(shape=[2, 3], dtype=paddle.float32, place=Place(gpu:0), stop_gradient=True,
             #       indices=[[0, 1, 2],
@@ -228,7 +228,7 @@ def sparse_csr_tensor(crows,
             cols = [1, 3, 2, 0, 1]
             values = [1, 2, 3, 4, 5]
             dense_shape = [3, 4]
-            csr = paddle.incubate.sparse.sparse_csr_tensor(crows, cols, values, dense_shape)
+            csr = paddle.sparse.sparse_csr_tensor(crows, cols, values, dense_shape)
             # print(csr)
             # Tensor(shape=[3, 4], dtype=paddle.int64, place=Place(gpu:0), stop_gradient=True,
             #       crows=[0, 2, 3, 5],
diff --git a/python/paddle/incubate/sparse/multiary.py b/python/paddle/sparse/multiary.py
similarity index 90%
rename from python/paddle/incubate/sparse/multiary.py
rename to python/paddle/sparse/multiary.py
index dd65b1bdee..32f12b97c2 100644
--- a/python/paddle/incubate/sparse/multiary.py
+++ b/python/paddle/sparse/multiary.py
@@ -58,6 +58,7 @@ def addmm(input, x, y, beta=1.0, alpha=1.0, name=None):
 
         .. code-block:: python
 
+            # required: gpu
             import paddle
 
             # dense + csr @ dense -> dense
@@ -65,17 +66,17 @@ def addmm(input, x, y, beta=1.0, alpha=1.0, name=None):
             crows = [0, 1, 2, 3]
             cols = [1, 2, 0]
             values = [1., 2., 3.]
-            x = paddle.incubate.sparse.sparse_csr_tensor(crows, cols, values, [3, 3])
+            x = paddle.sparse.sparse_csr_tensor(crows, cols, values, [3, 3])
             y = paddle.rand([3, 2])
-            out = paddle.incubate.sparse.addmm(input, x, y, 3.0, 2.0)
+            out = paddle.sparse.addmm(input, x, y, 3.0, 2.0)
 
             # dense + coo @ dense -> dense
             input = paddle.rand([3, 2])
             indices = [[0, 1, 2], [1, 2, 0]]
             values = [1., 2., 3.]
-            x = paddle.incubate.sparse.sparse_coo_tensor(indices, values, [3, 3])
+            x = paddle.sparse.sparse_coo_tensor(indices, values, [3, 3])
             y = paddle.rand([3, 2])
-            out = paddle.incubate.sparse.addmm(input, x, y, 3.0, 2.0)
+            out = paddle.sparse.addmm(input, x, y, 3.0, 2.0)
 
     """
     return _C_ops.sparse_addmm(input, x, y, alpha, beta)
diff --git a/python/paddle/incubate/sparse/nn/__init__.py b/python/paddle/sparse/nn/__init__.py
similarity index 100%
rename from python/paddle/incubate/sparse/nn/__init__.py
rename to python/paddle/sparse/nn/__init__.py
diff --git a/python/paddle/incubate/sparse/nn/functional/__init__.py b/python/paddle/sparse/nn/functional/__init__.py
similarity index 100%
rename from python/paddle/incubate/sparse/nn/functional/__init__.py
rename to python/paddle/sparse/nn/functional/__init__.py
diff --git a/python/paddle/incubate/sparse/nn/functional/activation.py b/python/paddle/sparse/nn/functional/activation.py
similarity index 95%
rename from python/paddle/incubate/sparse/nn/functional/activation.py
rename to python/paddle/sparse/nn/functional/activation.py
index 9e4db7e1c4..516cfc5cd5 100644
--- a/python/paddle/incubate/sparse/nn/functional/activation.py
+++ b/python/paddle/sparse/nn/functional/activation.py
@@ -43,7 +43,7 @@ def relu(x, name=None):
 
             dense_x = paddle.to_tensor([-2., 0., 1.])
             sparse_x = dense_x.to_sparse_coo(1)
-            out = paddle.incubate.sparse.nn.functional.relu(sparse_x)
+            out = paddle.sparse.nn.functional.relu(sparse_x)
             # [0., 0., 1.]
     """
     if in_dynamic_mode():
@@ -104,7 +104,7 @@ def softmax(x, axis=-1, name=None):
             #        values=[0.96823406, 0.19722934, 0.94373937, 0.02060066, 0.71456372,
             #                0.98275049])
 
-            out = paddle.incubate.sparse.nn.functional.softmax(csr)
+            out = paddle.sparse.nn.functional.softmax(csr)
             # Tensor(shape=[3, 4], dtype=paddle.float64, place=Place(gpu:0), stop_gradient=True,
             #        crows=[0, 2, 5, 6],
             #        cols=[2, 3, 0, 2, 3, 3],
@@ -139,7 +139,7 @@ def relu6(x, name=None):
 
             dense_x = paddle.to_tensor([-2., 0., 8.])
             sparse_x = dense_x.to_sparse_coo(1)
-            out = paddle.incubate.sparse.nn.functional.relu6(sparse_x)
+            out = paddle.sparse.nn.functional.relu6(sparse_x)
     """
     return _C_ops.sparse_relu6(x, 6.0)
 
@@ -175,6 +175,6 @@ def leaky_relu(x, negative_slope=0.01, name=None):
 
             dense_x = paddle.to_tensor([-2., 0., 5.])
             sparse_x = dense_x.to_sparse_coo(1)
-            out = paddle.incubate.sparse.nn.functional.leaky_relu(sparse_x, 0.5)
+            out = paddle.sparse.nn.functional.leaky_relu(sparse_x, 0.5)
     """
     return _C_ops.sparse_leaky_relu(x, negative_slope)
diff --git a/python/paddle/incubate/sparse/nn/functional/conv.py b/python/paddle/sparse/nn/functional/conv.py
similarity index 97%
rename from python/paddle/incubate/sparse/nn/functional/conv.py
rename to python/paddle/sparse/nn/functional/conv.py
index 233af394ec..03b0f2e9be 100644
--- a/python/paddle/incubate/sparse/nn/functional/conv.py
+++ b/python/paddle/sparse/nn/functional/conv.py
@@ -201,9 +201,9 @@ def conv3d(x,
               indices = paddle.to_tensor(indices, dtype='int32')
               values = paddle.to_tensor(values, dtype='float32')
               dense_shape = [1, 1, 3, 4, 1]
-              sparse_x = paddle.incubate.sparse.sparse_coo_tensor(indices, values, dense_shape, stop_gradient=True)
+              sparse_x = paddle.sparse.sparse_coo_tensor(indices, values, dense_shape, stop_gradient=True)
               weight = paddle.randn((1, 3, 3, 1, 1), dtype='float32')
-              y = paddle.incubate.sparse.nn.functional.conv3d(sparse_x, weight)
+              y = paddle.sparse.nn.functional.conv3d(sparse_x, weight)
               print(y.shape)
               # (1, 1, 1, 2, 1)
     """
@@ -321,9 +321,9 @@ def subm_conv3d(x,
               indices = paddle.to_tensor(indices, dtype='int32')
               values = paddle.to_tensor(values, dtype='float32')
               dense_shape = [1, 1, 3, 4, 1]
-              sparse_x = paddle.incubate.sparse.sparse_coo_tensor(indices, values, dense_shape, stop_gradient=True)
+              sparse_x = paddle.sparse.sparse_coo_tensor(indices, values, dense_shape, stop_gradient=True)
               weight = paddle.randn((1, 3, 3, 1, 1), dtype='float32')
-              y = paddle.incubate.sparse.nn.functional.subm_conv3d(sparse_x, weight)
+              y = paddle.sparse.nn.functional.subm_conv3d(sparse_x, weight)
               print(y.shape)
               #(1, 1, 3, 4, 1)
     """
diff --git a/python/paddle/incubate/sparse/nn/functional/pooling.py b/python/paddle/sparse/nn/functional/pooling.py
similarity index 97%
rename from python/paddle/incubate/sparse/nn/functional/pooling.py
rename to python/paddle/sparse/nn/functional/pooling.py
index aedadc8e56..d86ac58e92 100644
--- a/python/paddle/incubate/sparse/nn/functional/pooling.py
+++ b/python/paddle/sparse/nn/functional/pooling.py
@@ -70,7 +70,7 @@ def max_pool3d(x,
                 kernel_sizes = [3, 3, 3]
                 paddings = [0, 0, 0]
                 strides = [1, 1, 1]
-                out = paddle.incubate.sparse.nn.functional.max_pool3d(sparse_x, kernel_sizes, stride=strides, padding=paddings)
+                out = paddle.sparse.nn.functional.max_pool3d(sparse_x, kernel_sizes, stride=strides, padding=paddings)
                 #[1, 2, 2, 2, 3]
     """
 
diff --git a/python/paddle/incubate/sparse/nn/functional/transformer.py b/python/paddle/sparse/nn/functional/transformer.py
similarity index 95%
rename from python/paddle/incubate/sparse/nn/functional/transformer.py
rename to python/paddle/sparse/nn/functional/transformer.py
index 26164720cd..f17d728078 100644
--- a/python/paddle/incubate/sparse/nn/functional/transformer.py
+++ b/python/paddle/sparse/nn/functional/transformer.py
@@ -37,7 +37,7 @@ def attention(query,
 
     .. math::
 
-        result = softmax(\frac{ Q * K^T }{\sqrt{d}}) * V
+        result = softmax(\\frac{ Q * K^T }{\\sqrt{d}}) * V
 
     where : ``Q``, ``K``, and ``V`` represent the three input parameters of the attention module.
     The shape of the three parameters are: `[batch_size, num_heads, seq_len, head_dim]`, and
@@ -64,6 +64,7 @@ def attention(query,
     Examples:
         .. code-block:: python
 
+            # required: gpu
             import paddle
 
             batch_size = 16
@@ -85,7 +86,7 @@ def attention(query,
             kp_mask = paddle.randint(0, 2, [batch_size, seq_len])
             attn_mask = paddle.randint(0, 2, [seq_len, seq_len])
 
-            output = paddle.incubate.sparse.nn.functional.attention(query, key, value, sp_mask, kp_mask, attn_mask)
+            output = paddle.sparse.nn.functional.attention(query, key, value, sp_mask, kp_mask, attn_mask)
             output.backward()
     """
     return _C_ops.sparse_fused_attention(query, key, value, sparse_mask,
diff --git a/python/paddle/incubate/sparse/nn/layer/activation.py b/python/paddle/sparse/nn/layer/activation.py
similarity index 96%
rename from python/paddle/incubate/sparse/nn/layer/activation.py
rename to python/paddle/sparse/nn/layer/activation.py
index da5c5225ce..57c9d42dd5 100644
--- a/python/paddle/incubate/sparse/nn/layer/activation.py
+++ b/python/paddle/sparse/nn/layer/activation.py
@@ -41,7 +41,7 @@ class ReLU(Layer):
 
             dense_x = paddle.to_tensor([-2., 0., 1.])
             sparse_x = dense_x.to_sparse_coo(1)
-            relu = paddle.incubate.sparse.nn.ReLU()
+            relu = paddle.sparse.nn.ReLU()
             out = relu(sparse_x)
             # [0., 0., 1.]
     """
@@ -102,7 +102,7 @@ class Softmax(Layer):
             #        values=[0.96823406, 0.19722934, 0.94373937, 0.02060066, 0.71456372,
             #                0.98275049])
 
-            softmax = paddle.incubate.sparse.nn.Softmax()
+            softmax = paddle.sparse.nn.Softmax()
             out = softmax(csr)
             # Tensor(shape=[3, 4], dtype=paddle.float64, place=Place(gpu:0), stop_gradient=True,
             #        crows=[0, 2, 5, 6],
@@ -147,7 +147,7 @@ class ReLU6(Layer):
 
             dense_x = paddle.to_tensor([-2., 0., 8.])
             sparse_x = dense_x.to_sparse_coo(1)
-            relu6 = paddle.incubate.sparse.nn.ReLU6()
+            relu6 = paddle.sparse.nn.ReLU6()
             out = relu6(sparse_x)
     """
 
@@ -194,7 +194,7 @@ class LeakyReLU(Layer):
 
             dense_x = paddle.to_tensor([-2., 0., 5.])
             sparse_x = dense_x.to_sparse_coo(1)
-            leaky_relu = paddle.incubate.sparse.nn.LeakyReLU(0.5)
+            leaky_relu = paddle.sparse.nn.LeakyReLU(0.5)
             out = leaky_relu(sparse_x)
     """
 
diff --git a/python/paddle/incubate/sparse/nn/layer/conv.py b/python/paddle/sparse/nn/layer/conv.py
similarity index 98%
rename from python/paddle/incubate/sparse/nn/layer/conv.py
rename to python/paddle/sparse/nn/layer/conv.py
index 708359be16..8368eeeb03 100644
--- a/python/paddle/incubate/sparse/nn/layer/conv.py
+++ b/python/paddle/sparse/nn/layer/conv.py
@@ -216,8 +216,8 @@ class Conv3D(_Conv3D):
             indices = paddle.to_tensor(indices, dtype='int32')
             values = paddle.to_tensor(values, dtype='float32')
             dense_shape = [1, 1, 3, 4, 1]
-            sparse_x = paddle.incubate.sparse.sparse_coo_tensor(indices, values, dense_shape, stop_gradient=True)
-            conv = paddle.incubate.sparse.nn.Conv3D(1, 1, (1, 3, 3))
+            sparse_x = paddle.sparse.sparse_coo_tensor(indices, values, dense_shape, stop_gradient=True)
+            conv = paddle.sparse.nn.Conv3D(1, 1, (1, 3, 3))
             y = conv(sparse_x)
             print(y.shape)
             # (1, 1, 1, 2, 1)
@@ -353,8 +353,8 @@ class SubmConv3D(_Conv3D):
             dense_shape = [1, 1, 3, 4, 1]
             indices = paddle.to_tensor(indices, dtype='int32')
             values = paddle.to_tensor(values, dtype='float32')
-            sparse_x = paddle.incubate.sparse.sparse_coo_tensor(indices, values, dense_shape, stop_gradient=True)
-            subm_conv = paddle.incubate.sparse.nn.SubmConv3D(1, 1, (1, 3, 3))
+            sparse_x = paddle.sparse.sparse_coo_tensor(indices, values, dense_shape, stop_gradient=True)
+            subm_conv = paddle.sparse.nn.SubmConv3D(1, 1, (1, 3, 3))
             y = subm_conv(sparse_x)
             print(y.shape)
             # (1, 1, 3, 4, 1)
diff --git a/python/paddle/incubate/sparse/nn/layer/norm.py b/python/paddle/sparse/nn/layer/norm.py
similarity index 97%
rename from python/paddle/incubate/sparse/nn/layer/norm.py
rename to python/paddle/sparse/nn/layer/norm.py
index b84db710d7..b36a85f6b7 100644
--- a/python/paddle/incubate/sparse/nn/layer/norm.py
+++ b/python/paddle/sparse/nn/layer/norm.py
@@ -91,7 +91,7 @@ class BatchNorm(paddle.nn.BatchNorm1D):
               x_data = paddle.randn((1, 6, 6, 6, channels)).astype('float32')
               dense_x = paddle.to_tensor(x_data)
               sparse_x = dense_x.to_sparse_coo(4)
-              batch_norm = paddle.incubate.sparse.nn.BatchNorm(channels)
+              batch_norm = paddle.sparse.nn.BatchNorm(channels)
               batch_norm_out = batch_norm(sparse_x)
               print(batch_norm_out.shape)
               # [1, 6, 6, 6, 3]
@@ -262,7 +262,7 @@ class SyncBatchNorm(paddle.nn.SyncBatchNorm):
 
           # required: gpu
           import paddle
-          import paddle.incubate.sparse.nn as nn
+          import paddle.sparse.nn as nn
           import numpy as np
 
           x = np.array([[[[0.3, 0.4], [0.3, 0.07]], [[0.83, 0.37], [0.18, 0.93]]]]).astype('float32')
@@ -306,7 +306,7 @@ class SyncBatchNorm(paddle.nn.SyncBatchNorm):
     @classmethod
     def convert_sync_batchnorm(cls, layer):
         r"""
-        Helper function to convert :class: `paddle.incubate.sparse.nn.BatchNorm` layers in the model to :class: `paddle.incubate.sparse.nn.SyncBatchNorm` layers.
+        Helper function to convert :class: `paddle.sparse.nn.BatchNorm` layers in the model to :class: `paddle.sparse.nn.SyncBatchNorm` layers.
 
         Parameters:
             layer(paddle.nn.Layer): model containing one or more `BatchNorm` layers.
@@ -319,7 +319,7 @@ class SyncBatchNorm(paddle.nn.SyncBatchNorm):
             .. code-block:: python
 
                 import paddle
-                import paddle.incubate.sparse.nn as nn
+                import paddle.sparse.nn as nn
 
                 model = paddle.nn.Sequential(nn.Conv3D(3, 5, 3), nn.BatchNorm(5))
                 sync_model = nn.SyncBatchNorm.convert_sync_batchnorm(model)
diff --git a/python/paddle/incubate/sparse/nn/layer/pooling.py b/python/paddle/sparse/nn/layer/pooling.py
similarity index 98%
rename from python/paddle/incubate/sparse/nn/layer/pooling.py
rename to python/paddle/sparse/nn/layer/pooling.py
index 9fb67ecc0a..ee15bf6f79 100644
--- a/python/paddle/incubate/sparse/nn/layer/pooling.py
+++ b/python/paddle/sparse/nn/layer/pooling.py
@@ -66,7 +66,7 @@ class MaxPool3D(Layer):
             with _test_eager_guard():
                 dense_x = paddle.randn((2, 3, 6, 6, 3))
                 sparse_x = dense_x.to_sparse_coo(4)
-                max_pool3d = paddle.incubate.sparse.nn.MaxPool3D(
+                max_pool3d = paddle.sparse.nn.MaxPool3D(
                     kernel_size=3, data_format='NDHWC')
                 out = max_pool3d(sparse_x)
                 #shape=[2, 1, 2, 2, 3]
diff --git a/python/paddle/incubate/sparse/unary.py b/python/paddle/sparse/unary.py
similarity index 93%
rename from python/paddle/incubate/sparse/unary.py
rename to python/paddle/sparse/unary.py
index b7ba1536f1..c26ec35d3b 100644
--- a/python/paddle/incubate/sparse/unary.py
+++ b/python/paddle/sparse/unary.py
@@ -53,7 +53,7 @@ def sin(x, name=None):
 
             dense_x = paddle.to_tensor([-2., 0., 1.])
             sparse_x = dense_x.to_sparse_coo(1)
-            out = paddle.incubate.sparse.sin(sparse_x)
+            out = paddle.sparse.sin(sparse_x)
 
     """
     return _C_ops.sparse_sin(x)
@@ -83,7 +83,7 @@ def tan(x, name=None):
 
             dense_x = paddle.to_tensor([-2., 0., 1.])
             sparse_x = dense_x.to_sparse_coo(1)
-            out = paddle.incubate.sparse.tan(sparse_x)
+            out = paddle.sparse.tan(sparse_x)
 
     """
     return _C_ops.sparse_tan(x)
@@ -113,7 +113,7 @@ def asin(x, name=None):
 
             dense_x = paddle.to_tensor([-2., 0., 1.])
             sparse_x = dense_x.to_sparse_coo(1)
-            out = paddle.incubate.sparse.asin(sparse_x)
+            out = paddle.sparse.asin(sparse_x)
 
     """
     return _C_ops.sparse_asin(x)
@@ -144,7 +144,7 @@ def transpose(x, perm, name=None):
 
             dense_x = paddle.to_tensor([[-2., 0.], [1., 2.]])
             sparse_x = dense_x.to_sparse_coo(1)
-            out = paddle.incubate.sparse.transpose(sparse_x, [1, 0])
+            out = paddle.sparse.transpose(sparse_x, [1, 0])
 
     """
     return _C_ops.sparse_transpose(x, perm)
@@ -174,7 +174,7 @@ def atan(x, name=None):
 
             dense_x = paddle.to_tensor([-2., 0., 1.])
             sparse_x = dense_x.to_sparse_coo(1)
-            out = paddle.incubate.sparse.atan(sparse_x)
+            out = paddle.sparse.atan(sparse_x)
 
     """
     return _C_ops.sparse_atan(x)
@@ -204,7 +204,7 @@ def sinh(x, name=None):
 
             dense_x = paddle.to_tensor([-2., 0., 1.])
             sparse_x = dense_x.to_sparse_coo(1)
-            out = paddle.incubate.sparse.sinh(sparse_x)
+            out = paddle.sparse.sinh(sparse_x)
 
     """
     return _C_ops.sparse_sinh(x)
@@ -234,7 +234,7 @@ def asinh(x, name=None):
 
             dense_x = paddle.to_tensor([-2., 0., 1.])
             sparse_x = dense_x.to_sparse_coo(1)
-            out = paddle.incubate.sparse.asinh(sparse_x)
+            out = paddle.sparse.asinh(sparse_x)
 
     """
     return _C_ops.sparse_asinh(x)
@@ -264,7 +264,7 @@ def atanh(x, name=None):
 
             dense_x = paddle.to_tensor([-2., 0., 1.])
             sparse_x = dense_x.to_sparse_coo(1)
-            out = paddle.incubate.sparse.atanh(sparse_x)
+            out = paddle.sparse.atanh(sparse_x)
 
     """
     return _C_ops.sparse_atanh(x)
@@ -294,7 +294,7 @@ def tanh(x, name=None):
 
             dense_x = paddle.to_tensor([-2., 0., 1.])
             sparse_x = dense_x.to_sparse_coo(1)
-            out = paddle.incubate.sparse.tanh(sparse_x)
+            out = paddle.sparse.tanh(sparse_x)
 
     """
     return _C_ops.sparse_tanh(x)
@@ -324,7 +324,7 @@ def square(x, name=None):
 
             dense_x = paddle.to_tensor([-2., 0., 1.])
             sparse_x = dense_x.to_sparse_coo(1)
-            out = paddle.incubate.sparse.square(sparse_x)
+            out = paddle.sparse.square(sparse_x)
 
     """
     return _C_ops.sparse_square(x)
@@ -354,7 +354,7 @@ def sqrt(x, name=None):
 
             dense_x = paddle.to_tensor([-2., 0., 1.])
             sparse_x = dense_x.to_sparse_coo(1)
-            out = paddle.incubate.sparse.sqrt(sparse_x)
+            out = paddle.sparse.sqrt(sparse_x)
 
     """
     return _C_ops.sparse_sqrt(x)
@@ -384,7 +384,7 @@ def log1p(x, name=None):
 
             dense_x = paddle.to_tensor([-2, 0, 1], dtype='float32')
             sparse_x = dense_x.to_sparse_coo(1)
-            out = paddle.incubate.sparse.log1p(sparse_x)
+            out = paddle.sparse.log1p(sparse_x)
 
     """
     return _C_ops.sparse_log1p(x)
@@ -415,7 +415,7 @@ def cast(x, index_dtype=None, value_dtype=None, name=None):
 
             dense_x = paddle.to_tensor([-2, 0, 1])
             sparse_x = dense_x.to_sparse_coo(1)
-            out = paddle.incubate.sparse.cast(sparse_x, 'int32', 'float64')
+            out = paddle.sparse.cast(sparse_x, 'int32', 'float64')
 
     """
     if index_dtype and not isinstance(index_dtype, core.VarDesc.VarType):
@@ -450,7 +450,7 @@ def pow(x, factor, name=None):
 
             dense_x = paddle.to_tensor([-2, 0, 3], dtype='float32')
             sparse_x = dense_x.to_sparse_coo(1)
-            out = paddle.incubate.sparse.pow(sparse_x, 2)
+            out = paddle.sparse.pow(sparse_x, 2)
 
     """
     return _C_ops.sparse_pow(x, float(factor))
@@ -480,7 +480,7 @@ def neg(x, name=None):
 
             dense_x = paddle.to_tensor([-2, 0, 3], dtype='float32')
             sparse_x = dense_x.to_sparse_coo(1)
-            out = paddle.incubate.sparse.neg(sparse_x)
+            out = paddle.sparse.neg(sparse_x)
 
     """
     return _C_ops.sparse_scale(x, -1.0, 0.0, True)
@@ -510,7 +510,7 @@ def abs(x, name=None):
 
             dense_x = paddle.to_tensor([-2, 0, 3], dtype='float32')
             sparse_x = dense_x.to_sparse_coo(1)
-            out = paddle.incubate.sparse.abs(sparse_x)
+            out = paddle.sparse.abs(sparse_x)
 
     """
     return _C_ops.sparse_abs(x)
@@ -532,12 +532,10 @@ def coalesce(x):
 
             import paddle
 
-            from paddle.incubate import sparse
-
             indices = [[0, 0, 1], [1, 1, 2]]
             values = [1.0, 2.0, 3.0]
-            sp_x = sparse.sparse_coo_tensor(indices, values)
-            sp_x = sparse.coalesce(sp_x)
+            sp_x = paddle.sparse.sparse_coo_tensor(indices, values)
+            sp_x = paddle.sparse.coalesce(sp_x)
             print(sp_x.indices())
             #[[0, 1], [1, 2]]
             print(sp_x.values())
@@ -571,7 +569,7 @@ def rad2deg(x, name=None):
 
             dense_x = paddle.to_tensor([3.142, 0., -3.142])
             sparse_x = dense_x.to_sparse_coo(1)
-            out = paddle.incubate.sparse.rad2deg(sparse_x)
+            out = paddle.sparse.rad2deg(sparse_x)
 
     """
     if x.dtype in _int_dtype_:
@@ -604,7 +602,7 @@ def deg2rad(x, name=None):
 
             dense_x = paddle.to_tensor([-180, 0, 180])
             sparse_x = dense_x.to_sparse_coo(1)
-            out = paddle.incubate.sparse.deg2rad(sparse_x)
+            out = paddle.sparse.deg2rad(sparse_x)
 
     """
     if x.dtype in _int_dtype_:
@@ -636,7 +634,7 @@ def expm1(x, name=None):
 
             dense_x = paddle.to_tensor([-2., 0., 1.])
             sparse_x = dense_x.to_sparse_coo(1)
-            out = paddle.incubate.sparse.expm1(sparse_x)
+            out = paddle.sparse.expm1(sparse_x)
     """
     return _C_ops.sparse_expm1(x)
 
@@ -689,7 +687,7 @@ def reshape(x, shape, name=None):
                 sp_x = dense_x.to_sparse_coo(len(x_shape))
             else:
                 sp_x = dense_x.to_sparse_csr()
-            sp_out = paddle.incubate.sparse.reshape(sp_x, new_shape)
+            sp_out = paddle.sparse.reshape(sp_x, new_shape)
 
             print(sp_out)
             # the shape of sp_out is [1, 2, 2, 3, 3]
diff --git a/python/setup.py.in b/python/setup.py.in
index 648d7089e7..92b75cd067 100755
--- a/python/setup.py.in
+++ b/python/setup.py.in
@@ -283,7 +283,6 @@ packages=['paddle',
           'paddle.incubate.tensor',
           'paddle.incubate.multiprocessing',
           'paddle.incubate.nn',
-          'paddle.incubate.sparse',
           'paddle.incubate.asp',
           'paddle.incubate.passes',
           'paddle.distribution',
@@ -371,10 +370,10 @@ packages=['paddle',
           'paddle.vision.transforms',
           'paddle.vision.datasets',
           'paddle.audio',
-	  'paddle.audio.functional',
-	  'paddle.audio.features',
-	  'paddle.audio.datasets',
-	  'paddle.audio.backends',
+	      'paddle.audio.functional',
+	      'paddle.audio.features',
+	      'paddle.audio.datasets',
+	      'paddle.audio.backends',
           'paddle.text',
           'paddle.text.datasets',
           'paddle.incubate',
@@ -388,10 +387,10 @@ packages=['paddle',
           'paddle.incubate.distributed.models',
           'paddle.incubate.distributed.models.moe',
           'paddle.incubate.distributed.models.moe.gate',
-          'paddle.incubate.sparse',
-          'paddle.incubate.sparse.nn',
-          'paddle.incubate.sparse.nn.layer',
-          'paddle.incubate.sparse.nn.functional',
+          'paddle.sparse',
+          'paddle.sparse.nn',
+          'paddle.sparse.nn.layer',
+          'paddle.sparse.nn.functional',
           'paddle.incubate.xpu',
           'paddle.io',
           'paddle.optimizer',
diff --git a/tools/print_signatures.py b/tools/print_signatures.py
index f3c5f708b8..4c35fdf1c4 100644
--- a/tools/print_signatures.py
+++ b/tools/print_signatures.py
@@ -262,6 +262,9 @@ def check_public_api():
         paddle.audio,
         paddle.audio.backends,
         paddle.audio.datasets,
+        paddle.sparse,
+        paddle.sparse.nn,
+        paddle.sparse.nn.functional,
     ]
 
     apinum = 0
-- 
GitLab