!559 Add Fill/Erf op for VM.

Merge pull request !559 from liuxiao/fill-op

mindspore/ccsrc/kernel/tbe/tbe_adapter.cc

@@ -57,6 +57,7 @@ static std::map<string, string> tbe_func_adapter_map = {
   {"strided_slice", "strided_slice_d"},
   {"strided_slice_grad", "strided_slice_grad_d"},
   {"transpose", "transpose_d"},
+  {"fill", "fill_d"},
   {"unsorted_segment_sum", "unsorted_segment_sum_d"},
   {"concat", "concat_d"},
   {"slice", "slice_d"},

mindspore/ccsrc/pre_activate/pass/const_input_to_attr_registry.cc

@@ -53,6 +53,7 @@ ConstInputToAttrInfoRegistry::ConstInputToAttrInfoRegistry() {
   Register(kExpandDimsOpName, {1});
   Register(kSplitOpName, {0});
   Register(kTopKOpName, {1});
+  Register(kErfOpName, {1});
   Register(kSparseApplyAdagradOpName, {2});
   Register(kResizeNearestNeighborGrad, {1});
 }

mindspore/ccsrc/utils/utils.h

@@ -92,6 +92,7 @@ constexpr auto kClipByNormNoDivSumOpName = "ClipByNormNoDivSum";
 constexpr auto kGreaterOpName = "Greater";
 constexpr auto kSqrtOpName = "Sqrt";
 constexpr auto kRsqrtOpName = "Rsqrt";
+constexpr auto kErfOpName = "Erf";
 constexpr auto kRealDivOpName = "RealDiv";
 constexpr auto kLambUpdateWithLROpName = "LambUpdateWithLR";
 constexpr auto kLambNextMVWithDecayOpName = "LambNextMVWithDecay";

mindspore/ops/_grad/grad_math_ops.py

@@ -17,6 +17,7 @@
 
 
 from functools import reduce
+import numpy as np
 from .. import functional as F
 from .. import operations as P
 from ..operations import _grad_ops as G
@@ -333,6 +334,23 @@ def get_bprop_log(self):
     return bprop
 
 
+@bprop_getters.register(P.Erf)
+def get_bprop_erf(self):
+    """Grad definition for `Erf` operation."""
+    exp = P.Exp()
+    square = P.Square()
+    sqrt = P.Sqrt()
+    cast = P.Cast()
+    dtype = P.DType()
+
+    def bprop(x, out, dout):
+        half_root_pi = cast(2 / sqrt(F.scalar_to_tensor(np.pi)), dtype(x))
+        x_square = square(x)
+        dx = dout * half_root_pi * exp(-x_square)
+        return (dx,)
+    return bprop
+
+
 @bprop_getters.register(P.Pow)
 def get_bprop_pow(self):
     """Grad definition for `Pow` operation."""

mindspore/ops/_op_impl/tbe/__init__.py

@@ -139,6 +139,8 @@ from .smooth_l1_loss_grad import _smooth_l1_loss_grad_tbe
 from .fused_mul_add import _fused_mul_add_tbe
 from .fused_mul_add_n import _fused_mul_add_n_tbe
 from .fused_mul_apply_momentum import _fused_mul_apply_momentum_tbe
+from .fill_d import _fill_d_op_tbe
+from .erf import _erf_op_tbe
 from .depthwise_conv2d import _depthwise_conv2d_tbe
 from .depthwise_conv2d_backprop_filter import _depthwise_conv2d_backprop_filter_tbe
 from .depthwise_conv2d_backprop_input import _depthwise_conv2d_backprop_input_tbe

mindspore/ops/_op_impl/tbe/erf.py

@@ -0,0 +1,39 @@
+# 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.
+# ============================================================================
+
+"""Erf op"""
+from mindspore.ops.op_info_register import op_info_register, TBERegOp, DataType
+
+erf_op_info = TBERegOp("Erf") \
+    .fusion_type("ELEMWISE") \
+    .async_flag(False) \
+    .binfile_name("erf.so") \
+    .compute_cost(10) \
+    .kernel_name("erf") \
+    .partial_flag(True) \
+    .op_pattern("formatAgnostic") \
+    .input(0, "x", False, "required", "all") \
+    .output(0, "y", False, "required", "all") \
+    .dtype_format(DataType.F16_5HD, DataType.F16_5HD) \
+    .dtype_format(DataType.F16_Default, DataType.F16_Default) \
+    .dtype_format(DataType.F32_5HD, DataType.F32_5HD) \
+    .dtype_format(DataType.F32_Default, DataType.F32_Default) \
+    .get_op_info()
+
+
+@op_info_register(erf_op_info)
+def _erf_op_tbe():
+    """Erf TBE register"""
+    return

mindspore/ops/_op_impl/tbe/fill_d.py

@@ -0,0 +1,55 @@
+# 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.
+# ============================================================================
+
+"""FillD op"""
+from mindspore.ops.op_info_register import op_info_register, TBERegOp, DataType
+
+fill_d_op_info = TBERegOp("FillD") \
+    .fusion_type("ELEMWISE") \
+    .async_flag(False) \
+    .binfile_name("fill_d.so") \
+    .compute_cost(10) \
+    .kernel_name("fill_d") \
+    .partial_flag(True) \
+    .attr("dims", "required", "listInt", "all") \
+    .input(0, "value", False, "required", "all") \
+    .output(0, "y", False, "required", "all") \
+    .dtype_format(DataType.F16_5HD, DataType.F16_5HD) \
+    .dtype_format(DataType.F16_FracZ, DataType.F16_FracZ) \
+    .dtype_format(DataType.F16_C1HWNCoC0, DataType.F16_C1HWNCoC0) \
+    .dtype_format(DataType.F16_Default, DataType.F16_Default) \
+    .dtype_format(DataType.F32_5HD, DataType.F32_5HD) \
+    .dtype_format(DataType.F32_FracZ, DataType.F32_FracZ) \
+    .dtype_format(DataType.F32_C1HWNCoC0, DataType.F32_C1HWNCoC0) \
+    .dtype_format(DataType.F32_Default, DataType.F32_Default) \
+    .dtype_format(DataType.I32_5HD, DataType.I32_5HD) \
+    .dtype_format(DataType.I32_FracZ, DataType.I32_FracZ) \
+    .dtype_format(DataType.I32_C1HWNCoC0, DataType.I32_C1HWNCoC0) \
+    .dtype_format(DataType.I32_Default, DataType.I32_Default) \
+    .dtype_format(DataType.I8_5HD, DataType.I8_5HD) \
+    .dtype_format(DataType.I8_FracZ, DataType.I8_FracZ) \
+    .dtype_format(DataType.I8_C1HWNCoC0, DataType.I8_C1HWNCoC0) \
+    .dtype_format(DataType.I8_Default, DataType.I8_Default) \
+    .dtype_format(DataType.U8_5HD, DataType.U8_5HD) \
+    .dtype_format(DataType.U8_FracZ, DataType.U8_FracZ) \
+    .dtype_format(DataType.U8_C1HWNCoC0, DataType.U8_C1HWNCoC0) \
+    .dtype_format(DataType.U8_Default, DataType.U8_Default) \
+    .get_op_info()
+
+
+@op_info_register(fill_d_op_info)
+def _fill_d_op_tbe():
+    """FillD TBE register"""
+    return

mindspore/ops/operations/__init__.py

@@ -39,7 +39,7 @@ from .control_ops import ControlDepend, GeSwitch, Merge
 from .inner_ops import ScalarCast
 from .math_ops import (Abs, ACos, AddN, AssignAdd, AssignSub, Atan2, BatchMatMul,
                        ReduceMax, ReduceMin, ReduceMean, ReduceSum, ReduceAll, ReduceProd, CumProd,
-                       Cos, Div, Equal, EqualCount, Exp, Floor, FloorDiv, FloorMod, Acosh,
+                       Cos, Div, Equal, EqualCount, Exp, Erf, Floor, FloorDiv, FloorMod, Acosh,
                        Greater, GreaterEqual, Less, LessEqual, Log, LogicalAnd,
                        LogicalNot, LogicalOr, MatMul, Maximum,
                        Minimum, Mul, Neg, NMSWithMask, NotEqual,
@@ -139,6 +139,7 @@ __all__ = [
     'ReLU',
     'ReLU6',
     'Elu',
+    'Erf',
     'Sigmoid',
     'HSwish',
     'HSigmoid',

mindspore/ops/operations/math_ops.py

@@ -1007,6 +1007,36 @@ class Log(PrimitiveWithInfer):
         return x
 
 
+class Erf(PrimitiveWithInfer):
+    r"""
+    Computes the Gauss error function of `input_x` element-wise.
+
+    Inputs:
+        - **input_x** (Tensor) - The input tensor.
+
+    Outputs:
+        Tensor, has the same shape and dtype as the `input_x`.
+
+    Examples:
+        >>> input_x = Tensor(np.array([-1, 0, 1, 2, 3]), mindspore.float32)
+        >>> erf = P.Erf()
+        >>> erf(input_x)
+        [-0.8427168, 0., 0.8427168, 0.99530876, 0.99997765]
+    """
+
+    @prim_attr_register
+    def __init__(self):
+        """init Erf"""
+        self.init_prim_io_names(inputs=['x'], outputs=['y'])
+
+    def infer_shape(self, x_shape):
+        return x_shape
+
+    def infer_dtype(self, x_type):
+        validator.check_tensor_type_same({"x": x_type}, [mstype.float16, mstype.float32], self.name)
+        return x_type
+
+
 class Minimum(_MathBinaryOp):
     """
     Computes the element-wise minimum of input tensors.

tests/ut/python/ops/test_ops.py

@@ -250,6 +250,10 @@ test_case_math_ops = [
         'block': P.Exp(),
         'desc_inputs': [[2, 3]],
         'desc_bprop': [[2, 3]]}),
+    ('Erf', {
+        'block': P.Erf(),
+        'desc_inputs': [Tensor(np.array([-2, -1, 0, 1, 2]).astype(np.float16))],
+        'desc_bprop': [Tensor(np.array([-2, -1, 0, 1, 2]).astype(np.float16))]}),
     ('Floor', {
         'block': P.Floor(),
         'desc_inputs': [[2, 512, 56, 56]],