* -------------------------------------------------------------------------
* This file is part of the MultimodalSDK project.
* Copyright (c) 2025 Huawei Technologies Co.,Ltd.
*
* MultimodalSDK is licensed under Mulan PSL v2.
* You can use this software according to the terms and conditions of the Mulan PSL v2.
* You may obtain a copy of Mulan PSL v2 at:
*
* http://license.coscl.org.cn/MulanPSL2
*
* 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 FIT FOR A PARTICULAR PURPOSE.
* See the Mulan PSL v2 for more details.
* -------------------------------------------------------------------------
* Description: tensor file for python.
* Author: ACC SDK
* Create: 2025
* History: NA
*/
#include "PyTensor.h"
#include <iostream>
#include "Python.h"
#include "PyUtil.h"
#include "acc/tensor/Tensor.h"
#include "acc/tensor/TensorOps.h"
#include "acc/ErrorCode.h"
#include "acc/utils/LogImpl.h"
namespace PyAcc {
Tensor::Tensor() : tensor_(std::make_shared<Acc::Tensor>()) {}
Tensor Tensor::clone() const
{
Tensor tensor;
Acc::ErrorCode ret = tensor_->Clone(*tensor.GetTensorPtr());
if (ret != Acc::SUCCESS) {
Acc::LogError << "Tensor clone failed.";
throw std::runtime_error("Tensor clone failed.");
}
return tensor;
}
void Tensor::set_format(Acc::TensorFormat tensorFormat)
{
Acc::ErrorCode ret = tensor_->SetFormat(tensorFormat);
if (ret != Acc::SUCCESS) {
Acc::LogError << "Tensor set_format failed.";
throw std::runtime_error("Tensor set_format failed.");
}
}
std::vector<size_t> Tensor::Shape() const
{
return tensor_->Shape();
}
Acc::DataType Tensor::Dtype() const
{
return tensor_->DType();
}
const std::string& Tensor::Device()
{
std::unique_ptr<char[]> device = tensor_->Device();
if (!device) {
Acc::LogError << "Get device string failed, may be caused by out of memory, please check the memory status of "
"the environment.";
deviceStr_ = "";
} else {
std::string tmpStr(device.get());
deviceStr_ = tmpStr;
}
return deviceStr_;
}
Acc::TensorFormat Tensor::Format() const
{
return tensor_->Format();
}
size_t Tensor::NumBytes() const
{
return tensor_->NumBytes();
}
void Tensor::SetTensor(const Acc::Tensor& src)
{
(*tensor_) = src;
}
std::shared_ptr<Acc::Tensor> Tensor::GetTensorPtr() const
{
return tensor_;
}
Tensor Tensor::from_numpy(PyObject* pyObj)
{
NumpyData numpyData = GetNumpyData(pyObj);
Acc::Tensor accTensor(numpyData.dataPtr, numpyData.shape, numpyData.dataType, Acc::TensorFormat::ND, "cpu");
Tensor tensor;
tensor.SetTensor(accTensor);
return tensor;
}
PyObject* Tensor::numpy()
{
NumpyData numpyData;
numpyData.dataType = tensor_->DType();
numpyData.dataPtr = tensor_->Ptr();
numpyData.shape = tensor_->Shape();
try {
PyObject* numpyDataDict = ToNumpy(numpyData);
return numpyDataDict;
} catch (...) {
throw std::runtime_error("Failed to get __array_interface__ for numpy ndarray. "
"Maybe Python interpreter may be out of memory. ");
}
}
Tensor Tensor::normalize(const std::vector<float>& mean, const std::vector<float>& std,
const Acc::DeviceMode deviceMode)
{
Acc::Tensor outputAccTensor;
Acc::ErrorCode ret = Acc::TensorNormalize(*tensor_, outputAccTensor, mean, std, deviceMode);
if (ret != Acc::SUCCESS) {
throw std::runtime_error("Failed to execute normalize operator, please ensure your inputs are valid.");
}
Tensor tensor;
tensor.SetTensor(outputAccTensor);
return tensor;
}
}