Add ReluV2/ReluGradV2/ConfusionMulGrad for VM

mindspore/ccsrc/kernel/tbe/tbe_adapter.cc

@@ -33,6 +33,7 @@ static std::map<string, string> tbe_func_adapter_map = {
   {"re_lu6", "relu6"},
   {"re_lu6_grad", "relu6_grad"},
   {"re_lu", "relu"},
+  {"re_luv2", "relu_v2"},
   {"tensor_add", "add"},
   {"reduce_mean", "reduce_mean_d"},
   {"reduce_max", "reduce_max_d"},

mindspore/ops/_grad/grad_nn_ops.py

@@ -227,6 +227,18 @@ def get_bprop_relu6(self):
     return bprop
 
 
+@bprop_getters.register(P.ReLUV2)
+def get_bprop_relu_v2(self):
+    """Grad definition for `ReLUV2` operation."""
+    input_grad = G.ReluGradV2()
+
+    def bprop(x, out, dout):
+        mask = out[1]
+        dx = input_grad(dout[0], mask)
+        return (dx,)
+    return bprop
+
+
 @bprop_getters.register(P.HSwish)
 def get_bprop_hswish(self):
     """Grad definition for `HSwish` operation."""

mindspore/ops/_op_impl/tbe/__init__.py

@@ -33,6 +33,7 @@ from .cast import _cast_tbe
 from .conv2d import _conv2d_tbe
 from .conv2d_backprop_filter import _conv2d_backprop_filter_tbe
 from .conv2d_backprop_input import _conv2d_backprop_input_tbe
+from .confusion_mul_grad import _confusion_mul_grad_tbe
 from .dropout_do_mask import _dropout_do_mask_tbe
 from .gelu import _gelu_tbe
 from .gelu_grad import _gelu_grad_tbe
@@ -46,6 +47,8 @@ from .relu import _relu_tbe
 from .relu_grad import _relu_grad_tbe
 from .relu6 import _relu6_tbe
 from .relu6_grad import _relu6_grad_tbe
+from .relu_v2 import _relu_v2_tbe
+from .relu_grad_v2 import _relu_grad_v2_tbe
 from .softmax_cross_entropy_with_logits import _softmax_cross_entropy_with_logits_tbe
 from .sigmoid_cross_entropy_with_logits import _sigmoid_cross_entropy_with_logits_tbe
 from .sigmoid_cross_entropy_with_logits_grad import _sigmoid_cross_entropy_with_logits_grad_tbe

mindspore/ops/_op_impl/tbe/confusion_mul_grad.py

@@ -0,0 +1,38 @@
+# 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.
+# ============================================================================
+
+"""ConfusionMulGrad op"""
+from mindspore.ops.op_info_register import op_info_register, TBERegOp, DataType
+
+confusion_mul_grad_op_info = TBERegOp("ConfusionMulGrad") \
+    .fusion_type("OPAQUE") \
+    .attr("axis", "required", "listInt", "all") \
+    .attr("keep_dims", "required", "bool", "all") \
+    .input(0, "input0", False, "required", "all") \
+    .input(1, "input1", False, "required", "all") \
+    .input(2, "input2", False, "required", "all") \
+    .output(0, "output0", False, "required", "all") \
+    .output(1, "output1", False, "required", "all") \
+    .dtype_format(DataType.F16_Default, DataType.F16_Default, DataType.F16_Default,
+                  DataType.F16_Default, DataType.F16_Default) \
+    .dtype_format(DataType.F32_Default, DataType.F32_Default, DataType.F32_Default,
+                  DataType.F32_Default, DataType.F32_Default) \
+    .get_op_info()
+
+
+@op_info_register(confusion_mul_grad_op_info)
+def _confusion_mul_grad_tbe():
+    """ConfusionMulGrad TBE register"""
+    return

mindspore/ops/_op_impl/tbe/relu_grad_v2.py

@@ -0,0 +1,40 @@
+# 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.
+# ============================================================================
+
+"""ReluGradV2 op"""
+from mindspore.ops.op_info_register import op_info_register, TBERegOp, DataType
+
+relu_grad_v2_op_info = TBERegOp("ReluGradV2") \
+    .fusion_type("ELEMWISE") \
+    .async_flag(False) \
+    .binfile_name("relu_grad_v2.so") \
+    .compute_cost(10) \
+    .kernel_name("relu_grad_v2") \
+    .partial_flag(True) \
+    .input(0, "gradients", False, "required", "all") \
+    .input(1, "mask", False, "rerequired", "all") \
+    .output(0, "backprops", True, "required", "all") \
+    .dtype_format(DataType.F16_5HD, DataType.U8_Default, DataType.F16_5HD) \
+    .dtype_format(DataType.F32_5HD, DataType.U8_Default, DataType.F32_5HD) \
+    .dtype_format(DataType.I32_5HD, DataType.U8_Default, DataType.I32_5HD) \
+    .dtype_format(DataType.I8_5HD, DataType.U8_Default, DataType.I8_5HD) \
+    .dtype_format(DataType.U8_5HD, DataType.U8_Default, DataType.U8_5HD) \
+    .get_op_info()
+
+
+@op_info_register(relu_grad_v2_op_info)
+def _relu_grad_v2_tbe():
+    """ReluGradV2 TBE register"""
+    return

mindspore/ops/_op_impl/tbe/relu_v2.py

@@ -0,0 +1,40 @@
+# 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.
+# ============================================================================
+
+"""ReluV2 op"""
+from mindspore.ops.op_info_register import op_info_register, TBERegOp, DataType
+
+relu_v2_op_info = TBERegOp("ReLUV2") \
+    .fusion_type("ELEMWISE") \
+    .async_flag(False) \
+    .binfile_name("relu_v2.so") \
+    .compute_cost(10) \
+    .kernel_name("relu_v2") \
+    .partial_flag(True) \
+    .input(0, "x", False, "required", "all") \
+    .output(0, "y", False, "required", "all") \
+    .output(1, "mask", False, "required", "all") \
+    .dtype_format(DataType.F16_5HD, DataType.F16_5HD, DataType.U8_Default) \
+    .dtype_format(DataType.F32_5HD, DataType.F32_5HD, DataType.U8_Default) \
+    .dtype_format(DataType.I32_5HD, DataType.I32_5HD, DataType.U8_Default) \
+    .dtype_format(DataType.I8_5HD, DataType.I8_5HD, DataType.U8_Default) \
+    .dtype_format(DataType.U8_5HD, DataType.U8_5HD, DataType.U8_Default) \
+    .get_op_info()
+
+
+@op_info_register(relu_v2_op_info)
+def _relu_v2_tbe():
+    """ReluV2 TBE register"""
+    return

mindspore/ops/operations/__init__.py

@@ -58,8 +58,8 @@ from .nn_ops import (LSTM, SGD, Adam, ApplyMomentum, BatchNorm,
                      GetNext, L2Normalize, LayerNorm, L2Loss,
                      LogSoftmax,
                      MaxPool, ExtractImagePatches,
-                     AvgPool, Conv2DBackpropInput,
-                     MaxPoolWithArgmax, OneHot, Pad, MirrorPad, PReLU, ReLU, ReLU6, HSwish, HSigmoid,
+                     AvgPool, Conv2DBackpropInput, ConfusionMulGrad,
+                     MaxPoolWithArgmax, OneHot, Pad, MirrorPad, PReLU, ReLU, ReLU6, ReLUV2, HSwish, HSigmoid,
                      ResizeBilinear, Sigmoid,
                      SigmoidCrossEntropyWithLogits,
                      SmoothL1Loss, Softmax,
@@ -101,6 +101,7 @@ __all__ = [
     'LogSoftmax',
     'SoftmaxCrossEntropyWithLogits',
     'ROIAlign',
+    'ConfusionMulGrad',
     'SparseSoftmaxCrossEntropyWithLogits',
     'SGD',
     'ApplyMomentum',
@@ -138,6 +139,7 @@ __all__ = [
     'Split',
     'ReLU',
     'ReLU6',
+    'ReLUV2',
     'Elu',
     'Erf',
     'Sigmoid',

mindspore/ops/operations/_grad_ops.py

@@ -730,6 +730,27 @@ class ReLU6Grad(PrimitiveWithInfer):
         return x_dtype
 
 
+class ReluGradV2(PrimitiveWithInfer):
+    """Performs grad of ReLUV2 operation."""
+
+    @prim_attr_register
+    def __init__(self):
+        self.init_prim_io_names(inputs=['gradients', 'mask'], outputs=['output'])
+
+    def __call__(self, gradients, mask):
+        raise NotImplementedError
+
+    def infer_shape(self, gradients_shape, mask_shape):
+        return gradients_shape
+
+    def infer_dtype(self, gradients_dtype, mask_dtype):
+        args_type = {'gradients': gradients_dtype, 'mask': mask_dtype}
+        validator.check_args_tensor(args_type)
+        validator.check_typename("gradients_dtype", gradients_dtype, mstype.number_type)
+        validator.check_typename("mask_dtype", mask_dtype, (mstype.uint8,))
+        return gradients_dtype
+
+
 class EluGrad(PrimitiveWithInfer):
     """Performs grad of Elu operation."""
 

mindspore/ops/operations/array_ops.py

@@ -1329,7 +1329,7 @@ class Concat(PrimitiveWithInfer):
 
 def _get_pack_shape(x_shape, x_type, axis):
     """for pack output shape"""
-    validator.check_type("shape", x_shape, [tuple])
+    validator.check_type("shape", x_shape, [tuple, list])
     validator.check_integer("len of input_x shape", len(x_shape), 0, Rel.GT)
     validator.check_subclass("shape0", x_type[0], mstype.tensor)
     validator.check_integer("len of input_x0 shape", len(x_shape[0]), 0, Rel.GT)

mindspore/ops/operations/nn_ops.py

@@ -28,6 +28,7 @@ from ..._checkparam import Validator as validator
 from ..._checkparam import Rel
 from ...common import dtype as mstype
 from ..primitive import Primitive, PrimitiveWithInfer, prim_attr_register
+from ..operations.math_ops import _infer_shape_reduce
 
 
 def _check_positive_int_or_tuple(arg_name, arg_value, prim_name, allow_four=False, ret_four=False):
@@ -233,6 +234,62 @@ class ReLU6(PrimitiveWithInfer):
         return input_x
 
 
+class ReLUV2(PrimitiveWithInfer):
+    r"""
+    Computes ReLU(Rectified Linear Unit) of input tensor element-wise.
+
+    It returns :math:`\max(x,\  0)` element-wise.
+
+    Inputs:
+        - **input_x** (Tensor) - The input tensor should be a 4-D tensor.
+
+    Outputs:
+        - **output** (Tensor) - Has the same type and shape as the `input_x`.
+        - **mask** (Tensor) - A tensor whose data type must be uint8.
+
+    Examples:
+        >>> input_x = Tensor(np.array([[[[1, -2], [-3, 4]], [[-5, 6], [7, -8]]]]), mindspore.float32)
+        >>> relu_v2 = P.ReLUV2()
+        >>> output = relu_v2(input_x)
+        ([[[[1., 0.], [0., 4.]], [[0., 6.], [7., 0.]]]],
+         [[[[1, 0], [2, 0]], [[2, 0], [1, 0]]]])
+    """
+    @prim_attr_register
+    def __init__(self):
+        """init ReLUV2"""
+        self.init_prim_io_names(inputs=['x'], outputs=['output', 'mask'])
+
+    def __infer__(self, input_x):
+        input_shape = list(input_x['shape'])
+        input_dtype = input_x['dtype']
+        mask_shape = []
+        if len(input_shape) != 4:
+            raise ValueError("The `input_x` should be a 4-D tensor, "
+                             f"but got a {len(input_shape)}-D tensor whose shape is {input_shape}")
+        for i in enumerate(input_shape):
+            if i[0] == 1:
+                if input_dtype == mstype.uint8 and input_dtype == mstype.int8:
+                    mask_shape.append((input_shape[1] + 31) // 32)
+                else:
+                    mask_shape.append((input_shape[1] + 15) // 16)
+            else:
+                mask_shape.append(i[1])
+        if input_dtype == mstype.uint8 and input_dtype == mstype.int8:
+            mask_shape.append(4)
+        else:
+            mask_shape.append(2)
+
+        output_shape = (input_x['shape'], mask_shape)
+        validator.check_subclass("input_x", input_dtype, mstype.tensor, self.name)
+        validator.check_tensor_type_same({'input_x': input_dtype}, mstype.number_type, self.name)
+        mask_dtype = mstype.uint8
+        output_dtype = (input_dtype, mask_dtype)
+
+        return {'shape': output_shape,
+                'dtype': output_dtype,
+                'value': None}
+
+
 class Elu(PrimitiveWithInfer):
     r"""
     Computes exponential linear: `alpha * (exp(x) - 1)` if x < 0, `x` otherwise.
@@ -2580,3 +2637,51 @@ class ExtractImagePatches(PrimitiveWithInfer):
     def infer_dtype(self, input_x):
         validator.check_tensor_type_same({"input_x": input_x}, (mstype.int8, mstype.float16, mstype.float32), self.name)
         return input_x
+
+
+class ConfusionMulGrad(PrimitiveWithInfer):
+    """
+    `output0` is the result of which input0 dot multily input1.
+
+    `output1` is the result of which input0 dot multily input1, then reducesum it.
+
+    Args:
+        axis (Union[int, tuple[int], list[int]]): The dimensions to reduce.
+            Default:(), reduce all dimensions. Only constant value is allowed.
+        keep_dims (bool):
+            - If true, keep these reduced dimensions and the length is 1.
+            - If false, don't keep these dimensions. Default:False.
+
+    Inputs:
+        - **input_0** (Tensor) - The input Tensor.
+        - **input_1** (Tensor) - The input Tensor.
+        - **input_2** (Tensor) - The input Tensor.
+
+    outputs:
+        - **output_0** (Tensor) - The same shape with `input0`.
+        - **output_1** (Tensor)
+
+            - If axis is (), and keep_dims is false, the output is a 0-D array representing
+              the sum of all elements in the input array.
+            - If axis is int, set as 2, and keep_dims is false,
+              the shape of output is :math:`(x_1,x_3,...,x_R)`.
+            - If axis is tuple(int), set as (2,3), and keep_dims is false,
+              the shape of output is :math:`(x_1,x_4,...x_R)`.
+    """
+
+    @prim_attr_register
+    def __init__(self, axis = (), keep_dims = False):
+        self.init_prim_io_names(inputs = ["input0", "input1", "input2"], outputs = ["output0", "output1"])
+        self.axis_ = validator.check_value_type("axis", axis, [int, tuple, list], self.name)
+        self.keep_dims_ = validator.check_value_type("keep_dims", keep_dims, [bool], self.name)
+
+    def infer_shape(self, input0_shape, input1_shape, input2_shape):
+        outshape0 = input0_shape
+        outshape1 = _infer_shape_reduce(input1_shape, self.axis_, self.keep_dims_, self.name)
+        return outshape0, outshape1
+
+    def infer_dtype(self, input0_dtype, input1_dtype, input2_dtype):
+        validator.check_subclass("input0_dtype", input0_dtype, mstype.tensor, self.name)
+        validator.check_subclass("input1_dtype", input1_dtype, mstype.tensor, self.name)
+        validator.check_subclass("input2_dtype", input2_dtype, mstype.tensor, self.name)
+        return input0_dtype, input1_dtype

tests/st/ops/davinci/test_tbe_ops/test_relu_v2_grad.py

@@ -0,0 +1,53 @@
+# 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.
+# ============================================================================
+from mindspore import Tensor
+from mindspore.ops import operations as P
+import mindspore.nn as nn
+from mindspore.common.api import ms_function
+import numpy as np
+import mindspore.context as context
+from mindspore.common.initializer import initializer
+from mindspore.common.parameter import Parameter
+from mindspore.ops.composite import GradOperation
+context.set_context(mode=context.GRAPH_MODE, device_target="Ascend")
+
+class Grad(nn.Cell):
+    def __init__(self, network):
+        super(Grad, self).__init__()
+        self.grad = GradOperation(name="get_all", get_all=True)
+        self.network = network
+
+    @ms_function
+    def construct(self, input):
+        return self.grad(self.network)(input)
+
+class Net(nn.Cell):
+    def __init__(self):
+        super(Net, self).__init__()
+        self.relu_v2 = P.ReLUV2()
+
+    def construct(self, x):
+        return self.relu_v2(x)
+
+def test_net():
+    x = Tensor(np.ones((2,3,3,4)).astype(np.float32))
+    relu_net = Net()
+    relu_output = relu_net(x)
+    net = Grad(Net())
+    output_grad = net(x)
+    print(relu_output[0].asnumpy())
+    print(relu_output[1].asnumpy())
+    print(len(output_grad))
+    print(output_grad[0].asnumpy())

tests/ut/python/ops/test_ops.py

@@ -582,6 +582,10 @@ test_case_nn_ops = [
         'block': P.ReLU6(),
         'desc_inputs': [[1, 3, 4, 4]],
         'desc_bprop': [[1, 3, 4, 4]]}),
+    ('ReLUV2', {
+        'block': P.ReLUV2(),
+        'desc_inputs': [[1, 3, 4, 4]],
+        'desc_bprop': [[1, 3, 4, 4], [1, 3, 4, 4]]}),
     ('ReLUGrad', {
         'block': G.ReluGrad(),
         'desc_inputs': [[1, 3, 4, 4], [1, 3, 4, 4]],
@@ -1134,6 +1138,21 @@ test_case_other_ops = [
         'desc_inputs': [Tensor(np.array([1.1]).astype(np.float32)),
                         Tensor(np.array([1.2]).astype(np.float32))],
         'skip': ['backward']}),
+    ('ConfusionMulGrad_1', {
+        'block': P.ConfusionMulGrad(axis = [0], keep_dims = False),
+        'desc_inputs': [[3, 2], [3, 2], [3, 2]],
+        'desc_bprop': [[3, 2], [2]],
+        'skip': ['backward']}),
+    ('ConfusionMulGrad_2', {
+        'block': P.ConfusionMulGrad(axis = [0], keep_dims = True),
+        'desc_inputs': [[3, 2], [3, 2], [3, 2]],
+        'desc_bprop': [[3, 2], [1, 2]],
+        'skip': ['backward']}),
+    ('ConfusionMulGrad_3', {
+        'block':  P.ConfusionMulGrad(axis = (), keep_dims = True),
+        'desc_inputs': [[2, 3, 4], [2, 3, 4], [2, 3, 4]],
+        'desc_bprop': [[2, 3, 4], [1, 1, 1]],
+        'skip': ['backward']}),
     ('HistogramSummary', {
         'block': HistogramSummaryNet(),
         'desc_inputs': [Tensor(np.array([1.1]).astype(np.float32)),