* Copyright (c) 2025 Huawei Technologies Co., Ltd.
* This program is free software, you can redistribute it and/or modify it under the terms and conditions of
* CANN Open Software License Agreement Version 2.0 (the "License").
* Please refer to the License for details. You may not use this file except in compliance with the License.
* THIS SOFTWARE IS PROVIDED ON AN "AS IS" BASIS, WITHOUT WARRANTIES OF ANY KIND, EITHER EXPRESS OR IMPLIED,
* INCLUDING BUT NOT LIMITED TO NON-INFRINGEMENT, MERCHANTABILITY, OR FITNESS FOR A PARTICULAR PURPOSE.
* See LICENSE in the root of the software repository for the full text of the License.
*/
#include "onnx_common.h"
#include "math/reduce_sum/op_graph/reduce_sum_proto.h"
using namespace ge;
using ge::Operator;
namespace domi {
static Status parse_params_reduce_sum(const Message* op_src, ge::Operator& op_dest)
{
const ge::onnx::NodeProto* node = dynamic_cast<const ge::onnx::NodeProto*>(op_src);
if (node == nullptr) {
OP_LOGE(GetOpName(op_dest).c_str(), "Dynamic cast op_src to NodeProto failed.");
return FAILED;
}
int input_size = node->input_size();
int output_size = node->output_size();
op_dest.DynamicInputRegister("x", input_size);
op_dest.DynamicOutputRegister("output", output_size);
op_dest.SetAttr("original_type", "ai.onnx::11::ReduceSum");
std::vector<int> v_axes = {};
bool keep_dims_attr = true;
for (const auto& attr : node->attribute()) {
if (attr.name() == "axes" && attr.type() == ge::onnx::AttributeProto::INTS) {
for (int i = 0; i < attr.ints_size(); i++) {
v_axes.push_back(attr.ints(i));
}
} else if (attr.name() == "keepdims" && attr.type() == ge::onnx::AttributeProto::INT) {
keep_dims_attr = (attr.i() == 1);
}
}
int num = v_axes.size();
std::vector<int64_t> dims = {};
if (num != 0) {
dims.push_back(num);
}
ge::Tensor tensor = Vec2Tensor(v_axes, dims, ge::DT_INT32);
op_dest.SetAttr("axes", tensor);
op_dest.SetAttr("keep_dims", keep_dims_attr);
op_dest.SetAttr("name", node->name());
return SUCCESS;
}
static Status ParseOpToGraphReduceSum(const Operator& op, Graph& graph)
{
std::string ori_name;
if (op.GetAttr("name", ori_name) != SUCCESS) {
OP_LOGE(GetOpName(op).c_str(), "get name from op failed.");
return FAILED;
}
auto data0 = op::Data((ori_name + "_data0").c_str()).set_attr_index(0);
ge::Tensor value;
if (op.GetAttr("axes", value) != SUCCESS) {
OP_LOGE(GetOpName(op).c_str(), "get value from op failed");
return FAILED;
}
auto data1 = op::Const((ori_name + "_data1").c_str()).set_attr_value(value);
auto reducesum = op::ReduceSum((ori_name + "_ReduceSum").c_str()).set_input_x(data0).set_input_axes(data1);
bool flag = false;
if (op.GetAttr("keep_dims", flag) != SUCCESS) {
ge::AscendString op_name;
(void)op.GetName(op_name);
OP_LOGE(op_name.GetString(), "get keep_dims from op failed");
return FAILED;
}
reducesum.set_attr_keep_dims(flag);
std::vector<Operator> inputs{data0};
std::vector<std::pair<Operator, std::vector<size_t> > > output_indexs;
output_indexs.emplace_back(reducesum, vector<std::size_t>{0});
graph.SetInputs(inputs).SetOutputs(output_indexs);
return SUCCESS;
}
static Status ParseParamsReduceSum13(const Message* op_src, ge::Operator& op_dest)
{
const ge::onnx::NodeProto* node = dynamic_cast<const ge::onnx::NodeProto*>(op_src);
if (node == nullptr) {
OP_LOGE("ReduceSum13", "Dynamic cast op_src to NodeProto failed.");
return FAILED;
}
op_dest.SetAttr("original_type", "ai.onnx::13::ReduceSum");
int input_size = node->input_size();
bool keep_dims = true;
int noop_with_empty_axes = 0;
for (const auto& attr : node->attribute()) {
if (attr.name() == "keepdims" && attr.type() == ge::onnx::AttributeProto::INT) {
keep_dims = (attr.i() == 1);
} else if (attr.name() == "noop_with_empty_axes" && attr.type() == ge::onnx::AttributeProto::INT) {
noop_with_empty_axes = attr.i();
}
}
op_dest.SetAttr("name", node->name());
op_dest.SetAttr("input_size", input_size);
op_dest.SetAttr("keep_dims", keep_dims);
op_dest.SetAttr("noop_with_empty_axes", noop_with_empty_axes);
return SUCCESS;
}
namespace {
struct ReduceSum13Prop {
std::string ori_name;
bool keep_dims = false;
int input_num = 1;
int empty_axes = 0;
};
Status GetProperty(const Operator& op, ReduceSum13Prop& prop)
{
if (op.GetAttr("name", prop.ori_name) != SUCCESS) {
OP_LOGE(GetOpName(op).c_str(), "get name from op failed.");
return FAILED;
}
if (op.GetAttr("keep_dims", prop.keep_dims) != SUCCESS) {
OP_LOGE(GetOpName(op).c_str(), "get keep_dims from op failed");
return FAILED;
}
if (op.GetAttr("input_size", prop.input_num) != SUCCESS) {
OP_LOGE(GetOpName(op).c_str(), "get input_num from op failed");
return FAILED;
}
if (op.GetAttr("noop_with_empty_axes", prop.empty_axes) != SUCCESS) {
OP_LOGE(GetOpName(op).c_str(), "get attribute noop_with_empty_axes failed");
return FAILED;
}
return SUCCESS;
}
}
static Status GetInputTensorDimNum(const Operator& data_op, int64_t& dim_num) {
ge::TensorDesc input_desc = data_op.GetInputDesc(0);
auto shape = input_desc.GetShape();
if (shape.GetDimNum() <= 0) {
OP_LOGE("GetInputTensorDimNum", "Get input shape is invalid.");
return FAILED;
}
dim_num = shape.GetDimNum();
OP_LOGI(GetOpName(data_op).c_str(), "GetInputTensorDimNum is: %ld", dim_num);
return SUCCESS;
}
static Status ParseOpToGraphReduceSum13(const Operator& op, Graph& graph)
{
ReduceSum13Prop prop;
if (GetProperty(op, prop) != SUCCESS) {
return FAILED;
}
auto data0 = op::Data((prop.ori_name + "_data0").c_str()).set_attr_index(0);
int num_input = 2;
if (prop.input_num == 1 && prop.empty_axes == 0) {
int64_t input_dim_num = 0;
if (GetInputTensorDimNum(op, input_dim_num) != SUCCESS) {
OP_LOGE(GetOpName(op).c_str(), "Failed to get input tensor dimensions");
return FAILED;
}
std::vector<int64_t> v_axes;
for (int64_t i = 0; i < input_dim_num; ++i) {
v_axes.push_back(i);
}
ge::TensorDesc tensorDesc;
std::vector<int64_t> dims = {input_dim_num};
ge::Shape shape(dims);
tensorDesc.SetShape(shape);
tensorDesc.SetDataType(DT_INT64);
ge::Tensor tensor(tensorDesc, reinterpret_cast<uint8_t*>(v_axes.data()), v_axes.size() * sizeof(int64_t));
auto axes = op::Const((prop.ori_name + "_axes").c_str()).set_attr_value(tensor);
std::vector<Operator> inputs{data0, axes};
std::vector<std::pair<Operator, std::vector<size_t> > > output_indexs;
auto reducesum = op::ReduceSum((prop.ori_name + "_ReduceSum").c_str())
.set_input_x(data0)
.set_input_axes(axes)
.set_attr_keep_dims(prop.keep_dims)
.set_attr_noop_with_empty_axes(prop.empty_axes);
output_indexs.emplace_back(reducesum, vector<std::size_t>{0});
graph.SetInputs(inputs).SetOutputs(output_indexs);
} else if (prop.input_num == num_input) {
auto data1 = op::Data((prop.ori_name + "_data1").c_str()).set_attr_index(1);
auto reducesum13 = op::ReduceSum((prop.ori_name + "_ReduceSum").c_str())
.set_input_x(data0)
.set_input_axes(data1)
.set_attr_keep_dims(prop.keep_dims)
.set_attr_noop_with_empty_axes(prop.empty_axes);
std::vector<Operator> inputs{data0, data1};
std::vector<std::pair<Operator, std::vector<size_t> > > output_indexs;
output_indexs.emplace_back(reducesum13, vector<std::size_t>{0});
graph.SetInputs(inputs).SetOutputs(output_indexs);
} else {
OP_LOGE(GetOpName(op).c_str(), "Input num or set attr is error");
return FAILED;
}
return SUCCESS;
}
REGISTER_CUSTOM_OP("PartitionedCall")
.FrameworkType(ONNX)
.OriginOpType({ge::AscendString("ai.onnx::8::ReduceSum"),
ge::AscendString("ai.onnx::9::ReduceSum"),
ge::AscendString("ai.onnx::10::ReduceSum"),
ge::AscendString("ai.onnx::11::ReduceSum"),
ge::AscendString("ai.onnx::12::ReduceSum")})
.ParseParamsFn(parse_params_reduce_sum)
.ParseOpToGraphFn(ParseOpToGraphReduceSum)
.ImplyType(ImplyType::TVM);
REGISTER_CUSTOM_OP("ReduceSum")
.FrameworkType(ONNX)
.OriginOpType({ge::AscendString("ai.onnx::13::ReduceSum"),
ge::AscendString("ai.onnx::14::ReduceSum"),
ge::AscendString("ai.onnx::15::ReduceSum"),
ge::AscendString("ai.onnx::16::ReduceSum"),
ge::AscendString("ai.onnx::17::ReduceSum"),
ge::AscendString("ai.onnx::18::ReduceSum")})
.ParseParamsFn(ParseParamsReduceSum13)
.ParseOpToGraphFn(ParseOpToGraphReduceSum13)
.ImplyType(ImplyType::TVM);
}
