!10132 Add C.dot opt

From: @xutianming1985 Reviewed-by: Signed-off-by:
4 years ago · 5d97132f9e
parent f1bc935c11 d9830ddf63
commit 5d97132f9e
2 changed files with 257 additions and 0 deletions
--- a/mindspore/ops/composite/math_ops.py
+++ b/mindspore/ops/composite/math_ops.py
@ -260,3 +260,51 @@ def tensor_dot(x1, x2, axes):
    mul_result = matmul_op(x1_reshaped, x2_reshaped)
    final_result = reshape_op(mul_result, output_shape)
    return final_result
+
+
+@constexpr
+def _check_invalid_input(x1_shape, x2_shape):
+    if len(x1_shape) < 2 or len(x2_shape) < 2:
+        raise ValueError('C.dot inputs x1, x2 should has dimension >= 2,'
+                         + f'while x1 is ({len(x1_shape)}) and x2 is ({len(x2_shape)}).')
+
+
+def dot(x1, x2):
+    """
+    Computation a dot product between samples in two tensors.
+
+    Inputs:
+        - **x1** (Tensor) - First tensor in Dot op with datatype float16 or float32
+        - **x2** (Tensor) - Second tensor in Dot op with datatype float16 or float32
+
+    Outputs:
+        Tensor, dot product of x1 and x2.
+
+    Supported Platforms:
+        ``Ascend`` ``GPU`` ``CPU``
+
+    Examples:
+        >>> input_x1 = Tensor(np.ones(shape=[2, 3]), mindspore.float32)
+        >>> input_x2 = Tensor(np.ones(shape=[1, 3, 2]), mindspore.float32)
+        >>> output = C.Dot(input_x1, input_x2)
+        >>> print(output)
+        [[[3. 3.]]
+         [[3. 3.]]]
+    """
+    shape_op = P.Shape()
+    reshape_op = P.Reshape()
+    transpose_op = P.Transpose()
+    matmul_op = P.MatMul(False, False)
+    x1_shape = shape_op(x1)
+    x2_shape = shape_op(x2)
+    _check_invalid_input(x1_shape, x2_shape)
+
+    if len(x1_shape) > 2 or len(x2_shape) > 2:
+        x2_shape_range = range(len(x2_shape))
+        x2_shape_transpose = x2_shape_range[-2:-1] + x2_shape_range[:-2] + x2_shape_range[-1:]
+        x2_transpose = transpose_op(x2, x2_shape_transpose)
+        x1_reshape = reshape_op(x1, (-1, x1_shape[-1]))
+        x2_reshape = reshape_op(x2_transpose, (x2_shape[-2], -1))
+        mul_result = matmul_op(x1_reshape, x2_reshape)
+        return reshape_op(mul_result, x1_shape[:-1] + x2_shape[:-2] + x2_shape[-1:])
+    return matmul_op(x1, x2)
--- a/tests/st/ops/cpu/test_dot_op.py
+++ b/tests/st/ops/cpu/test_dot_op.py
@ -0,0 +1,209 @@
+# Copyright 2020 Huawei Technologies Co., Ltd
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+# http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+# ============================================================================
+
+import pytest
+import numpy as np
+from mindspore import Tensor
+import mindspore.nn as nn
+import mindspore.context as context
+from mindspore.ops import composite as C
+from mindspore.common.initializer import initializer
+
+
+context.set_context(mode=context.GRAPH_MODE, device_target="CPU")
+
+
+class NetDot(nn.Cell):
+    def construct(self, x, y):
+        return C.math_ops.dot(x, y)
+
+
+@pytest.mark.level0
+@pytest.mark.platform_x86_cpu
+@pytest.mark.env_onecard
+def test_dot_001():
+    x1_tensor = Tensor(np.array([[1., 2.], [4., 5.]]).astype(np.float32))
+    x2_tensor = Tensor(np.array([[[1., 2.], [3., 4.]], [[5., 6.], [7., 8.]], \
+                                 [[9., 10.], [11., 12.]]]).astype(np.float32))
+
+    network = NetDot()
+    ms_result_np = network(x1_tensor, x2_tensor)
+    expect_result = np.array([[[7., 10.], [19., 22.], [31., 34.]], \
+                              [[19., 28.], [55., 64.], [91., 100.]]]).astype(np.float32)
+    assert (ms_result_np.asnumpy() == expect_result).all()
+
+
+@pytest.mark.level0
+@pytest.mark.platform_x86_cpu
+@pytest.mark.env_onecard
+def test_dot_002():
+    x1_tensor = Tensor(np.array([[1., 2.], [4., 5.]]).astype(np.float32))
+    x2_tensor = Tensor(np.array([[[1., 2., 3.], [4., 5., 6.]], [[7., 8., 9.], [10., 11., 12.]]]).astype(np.float32))
+
+    network = NetDot()
+    ms_result_np = network(x1_tensor, x2_tensor)
+    expect_result = np.array([[[9., 12., 15.], [27., 30., 33.]], [[24., 33., 42.], [78., 87., 96.]]]).astype(np.float32)
+
+    assert (ms_result_np.asnumpy() == expect_result).all()
+
+
+@pytest.mark.level0
+@pytest.mark.platform_x86_cpu
+@pytest.mark.env_onecard
+def test_dot_003():
+    x1_tensor = initializer(Tensor(np.arange(2 * 3 * 4).reshape(2, 3, 4).astype(np.float32)), [2, 3, 4])
+    x2_tensor = initializer(Tensor(np.arange(1 * 5 * 4 * 2).reshape(1, 5, 4, 2).astype(np.float32)), [1, 5, 4, 2])
+
+    network = NetDot()
+    ms_result_np = network(x1_tensor, x2_tensor)
+    expect_result = np.array([[[[[28., 34.],
+                                 [76., 82.],
+                                 [124., 130.],
+                                 [172., 178.],
+                                 [220., 226.]]],
+                               [[[76., 98.],
+                                 [252., 274.],
+                                 [428., 450.],
+                                 [604., 626.],
+                                 [780., 802.]]],
+                               [[[124., 162.],
+                                 [428., 466.],
+                                 [732., 770.],
+                                 [1036., 1074.],
+                                 [1340., 1378.]]]],
+                              [[[[172., 226.],
+                                 [604., 658.],
+                                 [1036., 1090.],
+                                 [1468., 1522.],
+                                 [1900., 1954.]]],
+                               [[[220., 290.],
+                                 [780., 850.],
+                                 [1340., 1410.],
+                                 [1900., 1970.],
+                                 [2460., 2530.]]],
+                               [[[268., 354.],
+                                 [956., 1042.],
+                                 [1644., 1730.],
+                                 [2332., 2418.],
+                                 [3020., 3106.]]]]]).astype(np.float32)
+
+    assert (ms_result_np.asnumpy() == expect_result).all()
+
+
+@pytest.mark.level0
+@pytest.mark.platform_x86_cpu
+@pytest.mark.env_onecard
+def test_dot_004():
+    x1_tensor = initializer(Tensor(np.arange(3 * 4).reshape(3, 4).astype(np.float32)), [3, 4])
+    x2_tensor = initializer(Tensor(np.arange(4 * 5).reshape(4, 5).astype(np.float32)), [4, 5])
+
+    network = NetDot()
+    ms_result_np = network(x1_tensor, x2_tensor)
+    expect_result = np.array([[70., 76., 82., 88., 94.],
+                              [190., 212., 234., 256., 278.],
+                              [310., 348., 386., 424., 462.]]).astype(np.float32)
+
+    assert (ms_result_np.asnumpy() == expect_result).all()
+
+
+@pytest.mark.level0
+@pytest.mark.platform_x86_cpu
+@pytest.mark.env_onecard
+def test_dot_005():
+    x1_tensor = initializer(Tensor(np.arange(2 * 3 * 4).reshape(2, 3, 4).astype(np.float32)), [2, 3, 4])
+    x2_tensor = initializer(Tensor(np.arange(4 * 5).reshape(4, 5).astype(np.float32)), [4, 5])
+
+    network = NetDot()
+    ms_result_np = network(x1_tensor, x2_tensor)
+    expect_result = np.array([[[70., 76., 82., 88., 94.],
+                               [190., 212., 234., 256., 278.],
+                               [310., 348., 386., 424., 462.]],
+                              [[430., 484., 538., 592., 646.],
+                               [550., 620., 690., 760., 830.],
+                               [670., 756., 842., 928., 1014.]]]).astype(np.float32)
+
+    assert (ms_result_np.asnumpy() == expect_result).all()
+
+
+@pytest.mark.level0
+@pytest.mark.platform_x86_cpu
+@pytest.mark.env_onecard
+def test_dot_006():
+    x1_tensor = initializer(Tensor(np.arange(4).reshape(4).astype(np.float32)), [4])
+    x2_tensor = initializer(Tensor(np.arange(2 * 4 * 5).reshape(2, 4, 5).astype(np.float32)), [2, 4, 5])
+
+    network = NetDot()
+    try:
+        network(x1_tensor, x2_tensor)
+    except ValueError as e:
+        assert ValueError == type(e)
+
+
+def test_dot_007():
+    x1_tensor = initializer(Tensor(np.arange(4).reshape(4).astype(np.float32)), [4])
+    x2_tensor = initializer(Tensor(np.arange(4 * 4).reshape(4, 4).astype(np.float32)), [4, 4])
+
+    network = NetDot()
+    try:
+        network(x2_tensor, x1_tensor)
+    except ValueError as e:
+        assert ValueError == type(e)
+
+
+@pytest.mark.level0
+@pytest.mark.platform_x86_cpu
+@pytest.mark.env_onecard
+def test_dot_008():
+    x1_tensor = Tensor(np.array([]).astype(np.float32))
+    x2_tensor = Tensor(np.array([[[1., 2.], [3., 4.]],
+                                 [[5., 6.], [7., 8.]],
+                                 [[9., 10.], [11., 12.]]]).astype(np.float32))
+
+    network = NetDot()
+    try:
+        network(x2_tensor, x1_tensor)
+    except ValueError as e:
+        assert ValueError == type(e)
+
+
+@pytest.mark.level0
+@pytest.mark.platform_x86_cpu
+@pytest.mark.env_onecard
+def test_dot_009():
+    # for document
+    input_x1 = Tensor(np.array(np.ones(shape=[2, 3])).astype(np.float32))
+    input_x2 = Tensor(np.array(np.ones(shape=[1, 2, 3])).astype(np.float32))
+
+    network = NetDot()
+    try:
+        network(input_x1, input_x2)
+    except ValueError as e:
+        assert ValueError == type(e)
+
+
+@pytest.mark.level0
+@pytest.mark.platform_x86_cpu
+@pytest.mark.env_onecard
+def test_dot_010():
+    # for document
+    input_x1 = Tensor(np.array(np.ones(shape=[2, 3])).astype(np.float32))
+    input_x2 = Tensor(np.array(np.ones(shape=[1, 3, 2])).astype(np.float32))
+
+    network = NetDot()
+    ms_result_np = network(input_x1, input_x2)
+    expect_result = np.array([[[3., 3.]],
+                              [[3., 3.]]]).astype(np.float32)
+
+    assert (ms_result_np.asnumpy() == expect_result).all()