saymrwulf/onnxruntime
1
0
Fork 0
mirror of https://github.com/saymrwulf/onnxruntime.git synced 2026-05-14 20:48:00 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

Remove unnecessary const

Browse source
This commit is contained in:
Changming Sun 2018-12-13 15:52:04 -08:00
parent d6abc38182
commit afd831ae9e
5 changed files with 11 additions and 11 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

2
onnxruntime/core/framework/data_types.cc
View file

@ -726,7 +726,7 @@ const std::vector<MLDataType>& DataTypeImpl::AllTensorTypes() {
// helper to stream. expected to only be used for error output, so any typeid lookup
// cost should be fine. alternative would be to add a static string field to DataTypeImpl
// that we set in the register macro to the type name, and output that instead.
std::ostream& operator<<(std::ostream& out, const MLDataType data_type) {
std::ostream& operator<<(std::ostream& out, const DataTypeImpl* data_type) {
if (data_type == nullptr)
return out << "(null)";

12
onnxruntime/core/framework/execution_frame.cc
View file

@ -62,7 +62,7 @@ ExecutionFrame::ExecutionFrame(const std::unordered_map<std::string, MLValue>& f
ExecutionFrame::~ExecutionFrame() = default;
Status ExecutionFrame::AllocateMLValueTensorSelfOwnBuffer(int mlvalue_index,
const MLDataType element_type,
const DataTypeImpl* element_type,
const ONNXRuntimeAllocatorInfo& location,
const TensorShape& shape,
bool create_fence) {
@ -71,7 +71,7 @@ Status ExecutionFrame::AllocateMLValueTensorSelfOwnBuffer(int mlvalue_index,
}
Status ExecutionFrame::AllocateMLValueTensorSelfOwnBufferHelper(int mlvalue_index,
const MLDataType element_type,
const DataTypeImpl* element_type,
const ONNXRuntimeAllocatorInfo& location,
const TensorShape& shape,
bool create_fence) {
@ -163,7 +163,7 @@ void ExecutionFrame::TraceAllocate(int mlvalue_idx, size_t size) {
}
Status ExecutionFrame::AllocateTensorWithSelfOwnBuffer(const int index,
const MLDataType element_type,
const DataTypeImpl* element_type,
const ONNXRuntimeAllocatorInfo& location,
const TensorShape& shape,
bool create_fence) {
@ -173,7 +173,7 @@ Status ExecutionFrame::AllocateTensorWithSelfOwnBuffer(const int index,
Status ExecutionFrame::AllocateMLValueTensorPreAllocateBuffer(int mlvalue_index_to_allocate,
int mlvalue_index_reuse,
const MLDataType element_type,
const DataTypeImpl* element_type,
const ONNXRuntimeAllocatorInfo& location,
const TensorShape& shape,
bool create_fence) {
@ -200,7 +200,7 @@ Status ExecutionFrame::AllocateMLValueTensorPreAllocateBuffer(int mlvalue_index_
Status ExecutionFrame::AllocateTensorWithPreAllocateBufferHelper(MLValue* p_mlvalue,
void* pBuffer,
const MLDataType element_type,
const DataTypeImpl* element_type,
const ONNXRuntimeAllocatorInfo& location,
const TensorShape& shape) {
if (p_mlvalue->IsAllocated()) {
@ -219,7 +219,7 @@ Status ExecutionFrame::AllocateTensorWithPreAllocateBufferHelper(MLValue* p_mlva
Status ExecutionFrame::AllocateTensorWithPreAllocateBuffer(const int offset,
void* pBuffer,
const MLDataType element_type,
const DataTypeImpl* element_type,
const ONNXRuntimeAllocatorInfo& location,
const TensorShape& shape) {
ONNXRUNTIME_ENFORCE(offset >= 0 && offset < node_values_.size());

2
onnxruntime/core/framework/kernel_registry.cc
View file

@ -133,7 +133,7 @@ bool KernelRegistry::VerifyKernelDef(const onnxruntime::Node& node,
// valid names (of types or parameters) at the time that kernels are registered.
if ((nullptr != actual_type) &&
!std::any_of(allowed_types.begin(), allowed_types.end(),
[actual_type, &node, &error_str](const MLDataType& expected_type) {
[actual_type, &node, &error_str](const DataTypeImpl* expected_type) {
bool rc = expected_type->IsCompatible(*actual_type); // for easier debugging
if (!rc) {
// TODO print type information as well

4
onnxruntime/python/onnxruntime_pybind_mlvalue.cc
View file

@ -19,7 +19,7 @@ namespace python {
namespace py = pybind11;
using namespace onnxruntime::logging;
int OnnxRuntimeTensorToNumpyType(const MLDataType& tensor_type) {
int OnnxRuntimeTensorToNumpyType(const DataTypeImpl* tensor_type) {
static std::map<MLDataType, int> type_map{
{DataTypeImpl::GetType<bool>(), NPY_BOOL},
{DataTypeImpl::GetType<float>(), NPY_FLOAT},
@ -42,7 +42,7 @@ int OnnxRuntimeTensorToNumpyType(const MLDataType& tensor_type) {
}
}
const MLDataType& NumpyToOnnxRuntimeTensorType(int numpy_type) {
const DataTypeImpl* NumpyToOnnxRuntimeTensorType(int numpy_type) {
static std::map<int, MLDataType> type_map{
{NPY_BOOL, DataTypeImpl::GetType<bool>()},
{NPY_FLOAT, DataTypeImpl::GetType<float>()},

2
onnxruntime/python/onnxruntime_pybind_mlvalue.h
View file

@ -22,7 +22,7 @@ namespace python {
namespace py = pybind11;
int OnnxRuntimeTensorToNumpyType(const MLDataType& tensor_type);
int OnnxRuntimeTensorToNumpyType(const DataTypeImpl* tensor_type);
void CreateGenericMLValue(AllocatorPtr alloc, const std::string& name_input, py::object& value, MLValue* p_mlvalue);