* 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 "op_math_proto_extend.h"
#include "math/muls/op_graph/muls_proto.h"
#include "math/shape/op_graph/shape_proto.h"
using namespace ge;
namespace domi {
static Status ParseParamsRandomuniform(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;
}
const int input = 3;
const int output = 1;
op_dest.DynamicInputRegister("x", input);
op_dest.DynamicOutputRegister("y", output);
op_dest.SetAttr("original_type", "ai.onnx::11::RandomUniform");
float high = 1.0;
float low = 0.0;
std::vector<int> op_shape;
int seed = 0;
int dtype = 1;
for (const auto& attr : node->attribute()) {
if (attr.name() == "dtype" && attr.type() == ge::onnx::AttributeProto::INT) {
dtype = attr.i();
} else if (attr.name() == "high" && attr.type() == ge::onnx::AttributeProto::FLOAT) {
high = attr.f();
} else if (attr.name() == "low" && attr.type() == ge::onnx::AttributeProto::FLOAT) {
low = attr.f();
} else if (attr.name() == "seed" && attr.type() == ge::onnx::AttributeProto::FLOAT) {
seed = (int)attr.f();
} else if (attr.name() == "shape" && attr.type() == ge::onnx::AttributeProto::INTS) {
for (int i = 0; i < attr.ints_size(); i++) {
op_shape.push_back(attr.ints(i));
}
}
}
int num = op_shape.size();
std::vector<int64_t> dims = {num};
ge::Tensor tensor = Vec2Tensor(op_shape, dims, ge::DT_INT32);
op_dest.SetAttr("name", node->name());
op_dest.SetAttr("shape", tensor);
op_dest.SetAttr("max", high);
op_dest.SetAttr("min", low);
op_dest.SetAttr("seed", seed);
op_dest.SetAttr("dtype", dtype);
return SUCCESS;
}
namespace {
struct RandomuniformProp {
std::string ori_name;
float max_f = 0.0;
float min_f = 0.0;
int dtype = 1;
int seed = 0;
ge::Tensor shape;
};
Status GetProperty(const ge::Operator& op, RandomuniformProp& 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("shape", prop.shape) != SUCCESS) {
OP_LOGE(GetOpName(op).c_str(), "get shape from op failed");
return FAILED;
}
if (op.GetAttr("max", prop.max_f) != SUCCESS) {
OP_LOGE(GetOpName(op).c_str(), "get max from op failed");
return FAILED;
}
if (op.GetAttr("min", prop.min_f) != SUCCESS) {
OP_LOGE(GetOpName(op).c_str(), "get min from op failed");
return FAILED;
}
if (op.GetAttr("dtype", prop.dtype) != SUCCESS) {
OP_LOGE(GetOpName(op).c_str(), "get dtype from op failed");
return FAILED;
}
if (op.GetAttr("seed", prop.seed) != SUCCESS) {
OP_LOGE(GetOpName(op).c_str(), "get seed from op failed");
return FAILED;
}
return SUCCESS;
}
}
static Status ParseOpToGraphRandomuniform(const ge::Operator& op, ge::Graph& graph)
{
RandomuniformProp prop;
if (GetProperty(op, prop) != SUCCESS) {
return FAILED;
}
auto data0 = op::Const((prop.ori_name + "_data0").c_str()).set_attr_value(prop.shape);
std::map<int, int> kvlist = {{1, 0}, {10, 1}, {6, 3}, {2, 9}};
if (kvlist.find(prop.dtype) == kvlist.end()) {
OP_LOGE(GetOpName(op).c_str(), "only support float32/float16/int32/int64, but got %d", prop.dtype);
return FAILED;
}
ge::DataType temp_type = GetOmDtypeFromOnnxDtype(prop.dtype);
std::vector<std::pair<ge::Operator, std::vector<size_t>>> outputs;
if (static_cast<DataTypeOnnx>(prop.dtype) == DataTypeOnnx::DTO_INT32 ||
static_cast<DataTypeOnnx>(prop.dtype) == DataTypeOnnx::DTO_UINT8) {
int32_t max = prop.max_f;
ge::Tensor scalar_max = CreateScalar(max, temp_type);
auto data1 = op::Const((prop.ori_name + "_data1").c_str()).set_attr_value(scalar_max);
int32_t min = prop.min_f;
ge::Tensor scalar_min = CreateScalar(min, temp_type);
auto data2 = op::Const((prop.ori_name + "_data2").c_str()).set_attr_value(scalar_min);
auto random_int = op::RandomUniformInt((prop.ori_name + "_RandomUniformInt").c_str())
.set_input_shape(data0)
.set_input_min(data2)
.set_input_max(data1)
.set_attr_seed(prop.seed)
.set_attr_seed2(prop.seed);
std::vector<ge::Operator> inputs{data0, data1, data2};
outputs.emplace_back(random_int, std::vector<std::size_t>{0});
graph.SetInputs(inputs).SetOutputs(outputs);
} else {
auto random = op::RandomUniform((prop.ori_name + "_RandomUniform").c_str())
.set_input_shape(data0)
.set_attr_dtype(temp_type)
.set_attr_seed(prop.seed)
.set_attr_seed2(prop.seed);
float mul_num = prop.max_f - prop.min_f;
auto random_mul = op::Muls((prop.ori_name + "_Muls").c_str()).set_input_x(random).set_attr_value(mul_num);
auto random_float =
op::Adds((prop.ori_name + "_Adds").c_str()).set_input_x(random_mul).set_attr_value(prop.min_f);
std::vector<ge::Operator> inputs{data0};
outputs.emplace_back(random_float, std::vector<std::size_t>{0});
graph.SetInputs(inputs).SetOutputs(outputs);
}
return SUCCESS;
}
REGISTER_CUSTOM_OP("PartitionedCall")
.FrameworkType(ONNX)
.OriginOpType({ge::AscendString("ai.onnx::8::RandomUniform"),
ge::AscendString("ai.onnx::9::RandomUniform"),
ge::AscendString("ai.onnx::10::RandomUniform"),
ge::AscendString("ai.onnx::11::RandomUniform"),
ge::AscendString("ai.onnx::12::RandomUniform"),
ge::AscendString("ai.onnx::13::RandomUniform"),
ge::AscendString("ai.onnx::14::RandomUniform"),
ge::AscendString("ai.onnx::15::RandomUniform"),
ge::AscendString("ai.onnx::16::RandomUniform"),
ge::AscendString("ai.onnx::17::RandomUniform"),
ge::AscendString("ai.onnx::18::RandomUniform")})
.ParseParamsFn(ParseParamsRandomuniform)
.ParseOpToGraphFn(ParseOpToGraphRandomuniform)
.ImplyType(ImplyType::TVM);
}
