saymrwulf/onnxruntime
1
0
Fork 0
mirror of https://github.com/saymrwulf/onnxruntime.git synced 2026-07-30 20:18:08 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

Reduce header dependencies

Browse source
This commit is contained in:
Ryan Hill 2021-04-15 22:38:51 -07:00
parent 53dd09dbf2
commit 5205522c4f
3 changed files with 57 additions and 55 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

15
onnxruntime/core/framework/provider_bridge_ort.cc
View file

@ -726,6 +726,11 @@ struct ProviderHostImpl : ProviderHost {
int32_t Tensor__GetElementType(const Tensor* p) override { return p->GetElementType(); }
MLDataType Tensor__DataType(const Tensor* p) override { return p->DataType(); }
// AllocatorManager
void AllocatorManager__InsertAllocator(AllocatorManager* p, AllocatorPtr allocator) override { p->InsertAllocator(allocator); }
AllocatorPtr AllocatorManager__GetAllocator(const AllocatorManager* p, int id, OrtMemType mem_type) override { return p->GetAllocator(id, mem_type); };
#ifdef USE_CUDA
// GatherElements
Status GatherElements__ValidateInputShapes(const TensorShape& input_data_shape, const TensorShape& indices_shape, int64_t axis) override { return GatherElements::ValidateInputShapes(input_data_shape, indices_shape, axis); }
@ -812,10 +817,6 @@ struct ProviderHostImpl : ProviderHost {
Status EinsumTypedComputeProcessor__Run(EinsumTypedComputeProcessor<double>* p) override { return p->Run(); }
Status EinsumTypedComputeProcessor__Run(EinsumTypedComputeProcessor<MLFloat16>* p) override { return p->Run(); }
// AllocatorManager
void AllocatorManager__InsertAllocator(AllocatorManager* p, AllocatorPtr allocator) override { p->InsertAllocator(allocator); }
AllocatorPtr AllocatorManager__GetAllocator(const AllocatorManager* p, int id, OrtMemType mem_type) override { return p->GetAllocator(id, mem_type); };
std::unique_ptr<OpKernel> CreateOpKernel_CPU_If(const OpKernelInfo& info) override { return onnxruntime::make_unique<If>(info); }
std::unique_ptr<OpKernel> CreateOpKernel_CPU_Loop(const OpKernelInfo& info, const void* concat_output_func, void* stream) override { return Loop::Create(info, *reinterpret_cast<const Loop::ConcatOutput*>(concat_output_func), stream); }
std::unique_ptr<OpKernel> CreateOpKernel_CPU_Scan_8(const OpKernelInfo& info) override { return onnxruntime::make_unique<Scan<8>>(info); }
@ -825,9 +826,7 @@ struct ProviderHostImpl : ProviderHost {
#ifndef DISABLE_CONTRIB_OPS
Status embed_layer_norm__CheckInputs(const OpKernelContext* context) override { return contrib::embed_layer_norm::CheckInputs(context); }
Status bias_gelu_helper__CheckInputs(const OpKernelContext* context) override { return contrib::bias_gelu_helper::CheckInputs(context); }
// Status LongformerAttentionBase__CheckInputs(const contrib::LongformerAttentionBase* p, const TensorShape& input_shape, const TensorShape& weights_shape, const TensorShape& bias_shape, const TensorShape& mask_shape, const TensorShape& global_weights_shape, const TensorShape& global_bias_shape, const TensorShape& global_shape) override { return p->CheckInputs(input_shape, weights_shape, bias_shape, mask_shape, global_weights_shape, global_bias_shape, global_shape); }
Status LongformerAttentionBase__CheckInputs(const contrib::LongformerAttentionBase* p, const TensorShape& input_shape, const TensorShape& weights_shape, const TensorShape& bias_shape, const TensorShape& mask_shape, const TensorShape& global_weights_shape, const TensorShape& global_bias_shape, const TensorShape& global_shape) override { return Status::OK(); }
Status LongformerAttentionBase__CheckInputs(const contrib::LongformerAttentionBase* p, const TensorShape& input_shape, const TensorShape& weights_shape, const TensorShape& bias_shape, const TensorShape& mask_shape, const TensorShape& global_weights_shape, const TensorShape& global_bias_shape, const TensorShape& global_shape) override { return p->CheckInputs(input_shape, weights_shape, bias_shape, mask_shape, global_weights_shape, global_bias_shape, global_shape); }
Status AttentionBase__CheckInputs(const contrib::AttentionBase* p, const TensorShape& input_shape, const TensorShape& weights_shape, const TensorShape& bias_shape, const Tensor*& mask_index, const Tensor* past, const int max_threads_per_block) override { return p->CheckInputs(input_shape, weights_shape, bias_shape, mask_index, past, max_threads_per_block); }
Tensor* AttentionBase__GetPresent(const contrib::AttentionBase* p, OpKernelContext* context, const Tensor* past, int batch_size, int head_size, int sequence_length, int& past_sequence_length) override { return p->GetPresent(context, past, batch_size, head_size, sequence_length, past_sequence_length); }
#endif
@ -843,7 +842,7 @@ struct ProviderHostImpl : ProviderHost {
Status contrib__PrepareForTrainingCompute(const TensorShape& input_shape, int num_outputs, int64_t& axis, int& before_dims, int& after_dims_including_split_axis, int& after_dims_excluding_split, std::vector<int64_t>& split_sizes) override { return contrib::PrepareForTrainingCompute(input_shape, num_outputs, axis, before_dims, after_dims_including_split_axis, after_dims_excluding_split, split_sizes); }
Status YieldOp__Compute(const contrib::YieldOp* p, OpKernelContext* context) override { return p->Compute(context); }
#endif
#endif
} provider_host_;
struct ProviderSharedLibrary {

59
onnxruntime/core/providers/shared_library/provider_bridge_provider.cc
View file

@ -357,6 +357,35 @@ std::unique_ptr<OpKernelInfo> CopyOpKernelInfo(const OpKernelInfo& info) {
namespace onnxruntime {
namespace utils {
template <>
Status UnpackTensor(const ONNX_NAMESPACE::TensorProto& tensor, const void* raw_data, size_t raw_data_len, /*out*/ bool* p_data, size_t expected_size) { return g_host->UnpackTensor(tensor, raw_data, raw_data_len, p_data, expected_size); }
template <>
Status UnpackTensor(const ONNX_NAMESPACE::TensorProto& tensor, const void* raw_data, size_t raw_data_len, /*out*/ float* p_data, size_t expected_size) { return g_host->UnpackTensor(tensor, raw_data, raw_data_len, p_data, expected_size); }
template <>
Status UnpackTensor(const ONNX_NAMESPACE::TensorProto& tensor, const void* raw_data, size_t raw_data_len, /*out*/ double* p_data, size_t expected_size) { return g_host->UnpackTensor(tensor, raw_data, raw_data_len, p_data, expected_size); }
template <>
Status UnpackTensor(const ONNX_NAMESPACE::TensorProto& tensor, const void* raw_data, size_t raw_data_len, /*out*/ MLFloat16* p_data, size_t expected_size) { return g_host->UnpackTensor(tensor, raw_data, raw_data_len, p_data, expected_size); }
template <>
Status UnpackTensor(const ONNX_NAMESPACE::TensorProto& tensor, const void* raw_data, size_t raw_data_len, /*out*/ int8_t* p_data, size_t expected_size) { return g_host->UnpackTensor(tensor, raw_data, raw_data_len, p_data, expected_size); }
template <>
Status UnpackTensor(const ONNX_NAMESPACE::TensorProto& tensor, const void* raw_data, size_t raw_data_len, /*out*/ uint8_t* p_data, size_t expected_size) { return g_host->UnpackTensor(tensor, raw_data, raw_data_len, p_data, expected_size); }
template <>
Status UnpackTensor(const ONNX_NAMESPACE::TensorProto& tensor, const void* raw_data, size_t raw_data_len, /*out*/ int16_t* p_data, size_t expected_size) { return g_host->UnpackTensor(tensor, raw_data, raw_data_len, p_data, expected_size); }
template <>
Status UnpackTensor(const ONNX_NAMESPACE::TensorProto& tensor, const void* raw_data, size_t raw_data_len, /*out*/ uint16_t* p_data, size_t expected_size) { return g_host->UnpackTensor(tensor, raw_data, raw_data_len, p_data, expected_size); }
template <>
Status UnpackTensor(const ONNX_NAMESPACE::TensorProto& tensor, const void* raw_data, size_t raw_data_len, /*out*/ int32_t* p_data, size_t expected_size) { return g_host->UnpackTensor(tensor, raw_data, raw_data_len, p_data, expected_size); }
template <>
Status UnpackTensor(const ONNX_NAMESPACE::TensorProto& tensor, const void* raw_data, size_t raw_data_len, /*out*/ uint32_t* p_data, size_t expected_size) { return g_host->UnpackTensor(tensor, raw_data, raw_data_len, p_data, expected_size); }
template <>
Status UnpackTensor(const ONNX_NAMESPACE::TensorProto& tensor, const void* raw_data, size_t raw_data_len, /*out*/ int64_t* p_data, size_t expected_size) { return g_host->UnpackTensor(tensor, raw_data, raw_data_len, p_data, expected_size); }
template <>
Status UnpackTensor(const ONNX_NAMESPACE::TensorProto& tensor, const void* raw_data, size_t raw_data_len, /*out*/ uint64_t* p_data, size_t expected_size) { return g_host->UnpackTensor(tensor, raw_data, raw_data_len, p_data, expected_size); }
} // namespace utils
#ifdef USE_CUDA
bool TileOp::IsTileMemcpy(const TensorShape& input_shape, const int64_t* repeats, size_t rank, bool& is_batched_memcpy, size_t& num_of_elements_per_batch, size_t& num_of_copies_per_batch, size_t& num_of_batch_copies) {
return g_host->TileOp__IsTileMemcpy(input_shape, repeats, rank, is_batched_memcpy, num_of_elements_per_batch, num_of_copies_per_batch, num_of_batch_copies);
}
@ -411,34 +440,6 @@ std::unique_ptr<EinsumTypedComputeProcessor<float>> EinsumTypedComputeProcessor<
std::unique_ptr<EinsumTypedComputeProcessor<double>> EinsumTypedComputeProcessor<double>::Create(OpKernelContext* context, AllocatorPtr allocator, concurrency::ThreadPool* tp, EinsumComputePreprocessor& einsum_compute_preprocessor, void* einsum_cuda_assets) { return g_host->EinsumTypedComputeProcessor_double__Create(context, allocator, tp, einsum_compute_preprocessor, einsum_cuda_assets); }
std::unique_ptr<EinsumTypedComputeProcessor<MLFloat16>> EinsumTypedComputeProcessor<MLFloat16>::Create(OpKernelContext* context, AllocatorPtr allocator, concurrency::ThreadPool* tp, EinsumComputePreprocessor& einsum_compute_preprocessor, void* einsum_cuda_assets) { return g_host->EinsumTypedComputeProcessor_MLFloat16__Create(context, allocator, tp, einsum_compute_preprocessor, einsum_cuda_assets); }
namespace utils {
template <>
Status UnpackTensor(const ONNX_NAMESPACE::TensorProto& tensor, const void* raw_data, size_t raw_data_len, /*out*/ bool* p_data, size_t expected_size) { return g_host->UnpackTensor(tensor, raw_data, raw_data_len, p_data, expected_size); }
template <>
Status UnpackTensor(const ONNX_NAMESPACE::TensorProto& tensor, const void* raw_data, size_t raw_data_len, /*out*/ float* p_data, size_t expected_size) { return g_host->UnpackTensor(tensor, raw_data, raw_data_len, p_data, expected_size); }
template <>
Status UnpackTensor(const ONNX_NAMESPACE::TensorProto& tensor, const void* raw_data, size_t raw_data_len, /*out*/ double* p_data, size_t expected_size) { return g_host->UnpackTensor(tensor, raw_data, raw_data_len, p_data, expected_size); }
template <>
Status UnpackTensor(const ONNX_NAMESPACE::TensorProto& tensor, const void* raw_data, size_t raw_data_len, /*out*/ MLFloat16* p_data, size_t expected_size) { return g_host->UnpackTensor(tensor, raw_data, raw_data_len, p_data, expected_size); }
template <>
Status UnpackTensor(const ONNX_NAMESPACE::TensorProto& tensor, const void* raw_data, size_t raw_data_len, /*out*/ int8_t* p_data, size_t expected_size) { return g_host->UnpackTensor(tensor, raw_data, raw_data_len, p_data, expected_size); }
template <>
Status UnpackTensor(const ONNX_NAMESPACE::TensorProto& tensor, const void* raw_data, size_t raw_data_len, /*out*/ uint8_t* p_data, size_t expected_size) { return g_host->UnpackTensor(tensor, raw_data, raw_data_len, p_data, expected_size); }
template <>
Status UnpackTensor(const ONNX_NAMESPACE::TensorProto& tensor, const void* raw_data, size_t raw_data_len, /*out*/ int16_t* p_data, size_t expected_size) { return g_host->UnpackTensor(tensor, raw_data, raw_data_len, p_data, expected_size); }
template <>
Status UnpackTensor(const ONNX_NAMESPACE::TensorProto& tensor, const void* raw_data, size_t raw_data_len, /*out*/ uint16_t* p_data, size_t expected_size) { return g_host->UnpackTensor(tensor, raw_data, raw_data_len, p_data, expected_size); }
template <>
Status UnpackTensor(const ONNX_NAMESPACE::TensorProto& tensor, const void* raw_data, size_t raw_data_len, /*out*/ int32_t* p_data, size_t expected_size) { return g_host->UnpackTensor(tensor, raw_data, raw_data_len, p_data, expected_size); }
template <>
Status UnpackTensor(const ONNX_NAMESPACE::TensorProto& tensor, const void* raw_data, size_t raw_data_len, /*out*/ uint32_t* p_data, size_t expected_size) { return g_host->UnpackTensor(tensor, raw_data, raw_data_len, p_data, expected_size); }
template <>
Status UnpackTensor(const ONNX_NAMESPACE::TensorProto& tensor, const void* raw_data, size_t raw_data_len, /*out*/ int64_t* p_data, size_t expected_size) { return g_host->UnpackTensor(tensor, raw_data, raw_data_len, p_data, expected_size); }
template <>
Status UnpackTensor(const ONNX_NAMESPACE::TensorProto& tensor, const void* raw_data, size_t raw_data_len, /*out*/ uint64_t* p_data, size_t expected_size) { return g_host->UnpackTensor(tensor, raw_data, raw_data_len, p_data, expected_size); }
} // namespace utils
#ifndef DISABLE_CONTRIB_OPS
namespace contrib {
Status embed_layer_norm::CheckInputs(const OpKernelContext* context) { return g_host->embed_layer_norm__CheckInputs(context); }
@ -468,5 +469,5 @@ Status PassThrough::Compute(OpKernelContext* context) const { return g_host->con
Status YieldOp::Compute(OpKernelContext* context) const { return g_host->YieldOp__Compute(this, context); }
} // namespace contrib
#endif
#endif
} // namespace onnxruntime

38
onnxruntime/core/providers/shared_library/provider_interfaces.h
View file

@ -643,6 +643,11 @@ struct ProviderHost {
virtual int32_t Tensor__GetElementType(const Tensor* p) = 0;
virtual MLDataType Tensor__DataType(const Tensor* p) = 0;
// AllocatorManager
virtual void AllocatorManager__InsertAllocator(AllocatorManager* p, AllocatorPtr allocator) = 0;
virtual AllocatorPtr AllocatorManager__GetAllocator(const AllocatorManager* p, int id, OrtMemType mem_type) = 0;
#ifdef USE_CUDA
// GatherElements
virtual Status GatherElements__ValidateInputShapes(const TensorShape& input_data_shape, const TensorShape& indices_shape, int64_t axis) = 0;
@ -728,10 +733,6 @@ struct ProviderHost {
virtual Status EinsumTypedComputeProcessor__Run(EinsumTypedComputeProcessor<double>* p) = 0;
virtual Status EinsumTypedComputeProcessor__Run(EinsumTypedComputeProcessor<MLFloat16>* p) = 0;
// AllocatorManager
virtual void AllocatorManager__InsertAllocator(AllocatorManager* p, AllocatorPtr allocator) = 0;
virtual AllocatorPtr AllocatorManager__GetAllocator(const AllocatorManager* p, int id, OrtMemType mem_type) = 0;
virtual std::unique_ptr<OpKernel> CreateOpKernel_CPU_If(const OpKernelInfo& info) = 0;
virtual std::unique_ptr<OpKernel> CreateOpKernel_CPU_Loop(const OpKernelInfo& info, const void* concat_output_func, void* stream) = 0;
virtual std::unique_ptr<OpKernel> CreateOpKernel_CPU_Scan_8(const OpKernelInfo& info) = 0;
@ -742,7 +743,6 @@ struct ProviderHost {
virtual Status embed_layer_norm__CheckInputs(const OpKernelContext* context) = 0;
virtual Status bias_gelu_helper__CheckInputs(const OpKernelContext* context) = 0;
virtual Status LongformerAttentionBase__CheckInputs(const contrib::LongformerAttentionBase* p, const TensorShape& input_shape, const TensorShape& weights_shape, const TensorShape& bias_shape, const TensorShape& mask_shape, const TensorShape& global_weights_shape, const TensorShape& global_bias_shape, const TensorShape& global_shape) = 0;
virtual Status AttentionBase__CheckInputs(const contrib::AttentionBase* p, const TensorShape& input_shape, const TensorShape& weights_shape, const TensorShape& bias_shape, const Tensor*& mask_index, const Tensor* past, const int max_threads_per_block) = 0;
virtual Tensor* AttentionBase__GetPresent(const contrib::AttentionBase* p, OpKernelContext* context, const Tensor* past, int batch_size, int head_size, int sequence_length, int& past_sequence_length) = 0;
#endif
@ -758,6 +758,7 @@ struct ProviderHost {
virtual Status contrib__PrepareForTrainingCompute(const TensorShape& input_shape, int num_outputs, int64_t& axis, int& before_dims, int& after_dims_including_split_axis, int& after_dims_excluding_split, std::vector<int64_t>& split_sizes) = 0;
virtual Status YieldOp__Compute(const contrib::YieldOp* p, OpKernelContext* context) = 0;
#endif
#endif
};
extern ProviderHost* g_host;
@ -1612,18 +1613,7 @@ bool IsDataTypeString(MLDataType dt_type);
} // namespace utils
#ifdef ENABLE_TRAINING
namespace contrib {
inline void record_event_in_tensor(const Tensor& event_id_tensor) { return g_host->contrib__record_event_in_tensor(event_id_tensor); }
inline void wait_event_in_tensor(const Tensor& event_id_tensor) { return g_host->contrib__wait_event_in_tensor(event_id_tensor); }
inline void VerifyLogitWeightAndLabelShape(const TensorShape& logit_shape, const TensorShape& label_shape, const TensorShape* weight_shape) { g_host->contrib__VerifyLogitWeightAndLabelShape(logit_shape, label_shape, weight_shape); }
inline void GetNDCFromLogitAndLabelShape(const TensorShape& logit_shape, const TensorShape& label_shape, int64_t& N_D, int64_t& C) { g_host->contrib__GetNDCFromLogitAndLabelShape(logit_shape, label_shape, N_D, C); }
inline void GetPermutationAndShape(bool ncd_to_ndc, const TensorShape& tensor_shape, std::vector<int64_t>& new_shape, std::vector<size_t>& permutations) { g_host->contrib__GetPermutationAndShape(ncd_to_ndc, tensor_shape, new_shape, permutations); }
inline Status PrepareForTrainingCompute(const TensorShape& input_shape, int num_outputs, int64_t& axis, int& before_dims, int& after_dims_including_split_axis, int& after_dims_excluding_split, std::vector<int64_t>& split_sizes) { return g_host->contrib__PrepareForTrainingCompute(input_shape, num_outputs, axis, before_dims, after_dims_including_split_axis, after_dims_excluding_split, split_sizes); }
} // namespace contrib
#endif
#ifdef USE_CUDA
namespace GatherElements {
inline Status ValidateInputShapes(const TensorShape& input_data_shape,
const TensorShape& indices_shape,
@ -1676,6 +1666,18 @@ inline std::unique_ptr<OpKernel> CreateOpKernel_CPU_If(const OpKernelInfo& info)
inline std::unique_ptr<OpKernel> CreateOpKernel_CPU_Scan_8(const OpKernelInfo& info) { return g_host->CreateOpKernel_CPU_Scan_8(info); }
inline std::unique_ptr<OpKernel> CreateOpKernel_CPU_Scan_9(const OpKernelInfo& info) { return g_host->CreateOpKernel_CPU_Scan_9(info); }
#endif
#ifdef ENABLE_TRAINING
namespace contrib {
inline void record_event_in_tensor(const Tensor& event_id_tensor) { return g_host->contrib__record_event_in_tensor(event_id_tensor); }
inline void wait_event_in_tensor(const Tensor& event_id_tensor) { return g_host->contrib__wait_event_in_tensor(event_id_tensor); }
inline void VerifyLogitWeightAndLabelShape(const TensorShape& logit_shape, const TensorShape& label_shape, const TensorShape* weight_shape) { g_host->contrib__VerifyLogitWeightAndLabelShape(logit_shape, label_shape, weight_shape); }
inline void GetNDCFromLogitAndLabelShape(const TensorShape& logit_shape, const TensorShape& label_shape, int64_t& N_D, int64_t& C) { g_host->contrib__GetNDCFromLogitAndLabelShape(logit_shape, label_shape, N_D, C); }
inline void GetPermutationAndShape(bool ncd_to_ndc, const TensorShape& tensor_shape, std::vector<int64_t>& new_shape, std::vector<size_t>& permutations) { g_host->contrib__GetPermutationAndShape(ncd_to_ndc, tensor_shape, new_shape, permutations); }
inline Status PrepareForTrainingCompute(const TensorShape& input_shape, int num_outputs, int64_t& axis, int& before_dims, int& after_dims_including_split_axis, int& after_dims_excluding_split, std::vector<int64_t>& split_sizes) { return g_host->contrib__PrepareForTrainingCompute(input_shape, num_outputs, axis, before_dims, after_dims_including_split_axis, after_dims_excluding_split, split_sizes); }
} // namespace contrib
#endif // ENABLE_TRAINING
#endif // USE_CUDA
#endif // SHARED_PROVIDER
} // namespace onnxruntime