mirror of
https://github.com/saymrwulf/onnxruntime.git
synced 2026-05-26 22:35:43 +00:00
5350abe19d
this is a big PR. we are going to move it up to layer_dev , which is still a L3 so we are still safe to do work there agile. we are going to move this into the L3 so that ryan can start doing intergration testing. we will pause for a full code review and integration test result prior to going into the L2. >>>> raw comments from previous commits >>> * LearningModelSession is cleaned up to use the adapter, and parts of binding are. * moved everything in the winmladapter made it all nano-com using, WRL to construct objects in the ORT side. base interfaces for everythign for winml to call cleaned up a bunch of winml to use the base interfaces. * more pieces * GetData across the abi. * renamed some namepsace cleaned up OrtValue cleaned up Tensor cleaned up custom ops. everything *but* learnignmodel should be clean * make sure it's building. winml.dll is still a monolith.
102 lines
No EOL
3.8 KiB
C++
102 lines
No EOL
3.8 KiB
C++
// Copyright (c) Microsoft Corporation.
|
|
// Licensed under the MIT License.
|
|
|
|
#pragma once
|
|
|
|
#include "TensorBuffer.h"
|
|
#include "MLValueHelpers.h"
|
|
|
|
//
|
|
// the Tensor class is the actual object for CPU memory buffers.
|
|
// TensorBase contains one of these to represent the raw memory
|
|
// GetCpuResource() returns it
|
|
//
|
|
namespace Windows::AI::MachineLearning {
|
|
template <typename T>
|
|
class Tensor {
|
|
private:
|
|
using TensorBuffer = TensorBuffer<T>;
|
|
using TensorBufferPtr = typename TensorBuffer::TensorBufferPtr;
|
|
|
|
TensorBufferPtr m_buffer;
|
|
std::vector<int64_t> m_shape;
|
|
winrt::com_ptr<_winmla::IWinMLAdapter> adapter_;
|
|
|
|
|
|
public:
|
|
Tensor() = delete;
|
|
|
|
Tensor(
|
|
_winmla::IWinMLAdapter* adapter,
|
|
std::vector<int64_t> const& shape,
|
|
winrt::Windows::Storage::Streams::IBuffer buffer) : m_shape(shape),
|
|
m_buffer(
|
|
TensorBuffer::Create(
|
|
static_cast<uint32_t>(
|
|
std::accumulate(
|
|
std::begin(shape),
|
|
std::end(shape),
|
|
static_cast<int64_t>(1),
|
|
std::multiplies<int64_t>())),
|
|
buffer)) {
|
|
adapter_.copy_from(adapter);
|
|
}
|
|
|
|
Tensor(
|
|
_winmla::IWinMLAdapter* adapter,
|
|
std::vector<int64_t> const& shape) : m_shape(shape),
|
|
m_buffer(
|
|
TensorBuffer::Create(
|
|
static_cast<uint32_t>(
|
|
std::accumulate(
|
|
std::begin(shape),
|
|
std::end(shape),
|
|
static_cast<int64_t>(1),
|
|
std::multiplies<int64_t>())))) {
|
|
adapter_.copy_from(adapter);
|
|
}
|
|
|
|
Tensor(
|
|
_winmla::IWinMLAdapter* adapter,
|
|
std::vector<int64_t> const&& shape) : m_shape(std::move(shape)),
|
|
m_buffer(
|
|
TensorBuffer::Create(
|
|
static_cast<uint32_t>(
|
|
std::accumulate(
|
|
std::begin(shape),
|
|
std::end(shape),
|
|
static_cast<int64_t>(1),
|
|
std::multiplies<int64_t>())))) {
|
|
adapter_.copy_from(adapter);
|
|
}
|
|
|
|
auto size() const {
|
|
return m_buffer->Size();
|
|
}
|
|
|
|
auto buffer() {
|
|
return m_buffer->Buffer();
|
|
}
|
|
|
|
_winmla::IOrtValue* GetValue() {
|
|
// Get the data type
|
|
auto type = adapter_->GetTensorType(TensorKindFrom<T>::Type);
|
|
// create the ml value
|
|
winrt::com_ptr<_winmla::IOrtValue> value;
|
|
WINML_THROW_IF_FAILED(adapter_->CreateCPUMLValue(&m_shape, type, buffer().second, value.put()));
|
|
return value.detach();
|
|
}
|
|
|
|
void set(uint32_t size, const T* pData) {
|
|
m_buffer->Set(size, pData);
|
|
}
|
|
|
|
void set(std::vector<T>&& other) {
|
|
m_buffer->Set(other);
|
|
}
|
|
|
|
const std::vector<int64_t>& shape() const {
|
|
return m_shape;
|
|
}
|
|
};
|
|
} // namespace Windows::AI::MachineLearning
|