add grad definition for AsinGrad and ACosGrad

4 years ago · 677127ef5e
parent 6dd5d2effd
commit 677127ef5e
3 changed files with 174 additions and 0 deletions
--- a/mindspore/ops/_grad/grad_math_ops.py
+++ b/mindspore/ops/_grad/grad_math_ops.py
@ -938,6 +938,19 @@ def get_bprop_asin(self):
    return bprop


+@bprop_getters.register(G.AsinGrad)
+def get_bprop_asin_grad(self):
+    """Grad definition for `AsinGrad` operation."""
+    input_grad = G.AsinGrad()
+    p_pow = P.Pow()
+
+    def bprop(x, grad, out, dout):
+        d2x = dout * grad * x * p_pow((1 - x * x), - 1.5)
+        ddy = input_grad(x, dout)
+        return (d2x, ddy)
+    return bprop
+
+
@bprop_getters.register(P.Asinh)
 def get_bprop_asinh(self):
    """Grad definition for `Asinh` operation."""
@ -986,6 +999,21 @@ def get_bprop_acos(self):
    return bprop


+@bprop_getters.register(G.ACosGrad)
+def get_bprop_acos_grad(self):
+    """Grad definition for `ACosGrad` operation."""
+    input_grad = G.ACosGrad()
+    p_pow = P.Pow()
+
+    def bprop(x, grad, out, dout):
+        d2x = -dout * grad * x * p_pow((1 - x * x), - 1.5)
+        ddy = input_grad(x, dout)
+        return (d2x, ddy)
+
+    return bprop
+
+
+
@bprop_getters.register(P.Acosh)
 def get_bprop_acosh(self):
    """Grad definition for `Acosh` operation."""
--- a/tests/st/ops/gpu/test_acos_grad_grad_op.py
+++ b/tests/st/ops/gpu/test_acos_grad_grad_op.py
@ -0,0 +1,73 @@
+# 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 numpy as np
+import pytest
+
+import mindspore.context as context
+import mindspore.nn as nn
+from mindspore import Tensor
+from mindspore.common.api import ms_function
+from mindspore.ops.operations import _grad_ops as G
+from mindspore.ops.composite import GradOperation
+
+
+class NetAcosGrad(nn.Cell):
+    def __init__(self):
+        super(NetAcosGrad, self).__init__()
+        self.acos_grad = G.ACosGrad()
+
+    @ms_function
+    def construct(self, x, dy):
+        return self.acos_grad(x, dy)
+
+
+class Grad(nn.Cell):
+    def __init__(self, network):
+        super(Grad, self).__init__()
+        self.grad = GradOperation(get_all=True, sens_param=True)
+        self.network = network
+
+    @ms_function
+    def construct(self, x, grad, dout):
+        return self.grad(self.network)(x, grad, dout)
+
+
+@pytest.mark.level0
+@pytest.mark.platform_x86_gpu_training
+@pytest.mark.env_onecard
+@pytest.mark.parametrize("fp_type, error_magnitude, mode", [
+    (np.float16, 1.0e-3, context.PYNATIVE_MODE),
+    (np.float32, 1.0e-6, context.PYNATIVE_MODE),
+    (np.float16, 1.0e-3, context.GRAPH_MODE),
+    (np.float32, 1.0e-6, context.GRAPH_MODE)
+])
+def test_acos_grad_grad(fp_type, error_magnitude, mode):
+    x = Tensor(np.array([0, -0.25, 0.5, 0.3]).astype(fp_type))
+    grad = Tensor(np.array([0, -0.25, 0.5, 0.3]).astype(fp_type))
+    dout = Tensor(np.array([2, 2, 2, 2]).astype(fp_type))
+
+    expect_ddy = np.array([-2, -2.0655911, -2.3094011, -2.0965697]).astype(fp_type)
+    expect_d2x = np.array([0, -0.1377061, -0.7698004, -0.2073530]).astype(fp_type)
+
+    error = np.ones(4) * error_magnitude
+
+    context.set_context(mode=mode, device_target="GPU")
+    acos_grad_grad = Grad(NetAcosGrad())
+    d2x, ddy = acos_grad_grad(x, grad, dout)
+    diff0 = ddy.asnumpy() - expect_ddy
+    diff1 = d2x.asnumpy() - expect_d2x
+    assert np.all(abs(diff0) < error)
+    assert np.all(abs(diff1) < error)
--- a/tests/st/ops/gpu/test_asin_grad_grad_op.py
+++ b/tests/st/ops/gpu/test_asin_grad_grad_op.py
@ -0,0 +1,73 @@
+# 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 numpy as np
+import pytest
+
+import mindspore.context as context
+import mindspore.nn as nn
+from mindspore import Tensor
+from mindspore.common.api import ms_function
+from mindspore.ops.operations import _grad_ops as G
+from mindspore.ops.composite import GradOperation
+
+
+class NetAsinGrad(nn.Cell):
+    def __init__(self):
+        super(NetAsinGrad, self).__init__()
+        self.asin_grad = G.AsinGrad()
+
+    @ms_function
+    def construct(self, x, dy):
+        return self.asin_grad(x, dy)
+
+
+class Grad(nn.Cell):
+    def __init__(self, network):
+        super(Grad, self).__init__()
+        self.grad = GradOperation(get_all=True, sens_param=True)
+        self.network = network
+
+    @ms_function
+    def construct(self, x, grad, dout):
+        return self.grad(self.network)(x, grad, dout)
+
+
+@pytest.mark.level0
+@pytest.mark.platform_x86_gpu_training
+@pytest.mark.env_onecard
+@pytest.mark.parametrize("fp_type, error_magnitude, mode", [
+    (np.float16, 1.0e-3, context.PYNATIVE_MODE),
+    (np.float32, 1.0e-6, context.PYNATIVE_MODE),
+    (np.float16, 1.0e-3, context.GRAPH_MODE),
+    (np.float32, 1.0e-6, context.GRAPH_MODE)
+])
+def test_asin_grad_grad(fp_type, error_magnitude, mode):
+    x = Tensor(np.array([0, -0.25, 0.5, 0.3]).astype(fp_type))
+    grad = Tensor(np.array([0, -0.25, 0.5, 0.3]).astype(fp_type))
+    dout = Tensor(np.array([2, 2, 2, 2]).astype(fp_type))
+
+    expect_ddy = np.array([2, 2.0655911, 2.3094011, 2.0965697]).astype(fp_type)
+    expect_d2x = np.array([0, 0.1377061, 0.7698004, 0.2073530]).astype(fp_type)
+
+    error = np.ones(4) * error_magnitude
+
+    context.set_context(mode=mode, device_target="GPU")
+    asin_grad_grad = Grad(NetAsinGrad())
+    d2x, ddy = asin_grad_grad(x, grad, dout)
+    diff0 = ddy.asnumpy() - expect_ddy
+    diff1 = d2x.asnumpy() - expect_d2x
+    assert np.all(abs(diff0) < error)
+    assert np.all(abs(diff1) < error)