ONNX Runtime
|
▼NOrt | All C++ Onnxruntime APIs are defined inside this namespace |
▼Ndetail | |
CAllocatedFree | |
CBase | Used internally by the C++ API. C++ wrapper types inherit from this. This is a zero cost abstraction to wrap the C API objects and delete them on destruction |
CBase< Unowned< T > > | Covers unowned pointers owned by either the ORT or some other instance of CPP wrappers. Used for ConstXXX and UnownedXXXX types that are copyable. Also convenient to wrap raw OrtXX pointers |
CConstIoBindingImpl | |
CConstSessionImpl | |
CConstValueImpl | |
CIoBindingImpl | |
CMapTypeInfoImpl | |
CMemoryInfoImpl | |
COptionalTypeInfoImpl | |
COrtSparseValuesParam | |
CSequenceTypeInfoImpl | |
CSessionImpl | |
CShape | |
CTensorTypeAndShapeInfoImpl | |
CTypeInfoImpl | |
CUnowned | This is a tagging template type. Use it with Base<T> to indicate that the C++ interface object has no ownership of the underlying C object |
CValueImpl | |
CAllocator | Wrapper around OrtAllocator |
CAllocatorWithDefaultOptions | Wrapper around OrtAllocator default instance that is owned by Onnxruntime |
CArenaCfg | It is a structure that represents the configuration of an arena based allocator |
CBFloat16_t | Bfloat16 (Brain Floating Point) data type |
CCheckpointState | Holds the state of the training session |
CCustomOpBase | |
CCustomOpConfigs | Class that represents session configuration entries for one or more custom operators |
CCustomOpDomain | Custom Op Domain |
CEnv | The Env (Environment) |
CException | All C++ methods that can fail will throw an exception of this type |
CFloat16_t | IEEE 754 half-precision floating point data type |
CFloat8E4M3FN_t | Float8e4m3fn (Float8 Floating Point) data type |
CFloat8E4M3FNUZ_t | Float8e4m3fnuz (Float8 Floating Point) data type |
CFloat8E5M2_t | Float8e5m2 (Float8 Floating Point) data type |
CFloat8E5M2FNUZ_t | Float8e5m2fnuz (Float8 Floating Point) data type |
CGlobal | |
CIoBinding | Wrapper around OrtIoBinding |
CKernelContext | This class wraps a raw pointer OrtKernelContext* that is being passed to the custom kernel Compute() method. Use it to safely access context attributes, input and output parameters with exception safety guarantees. See usage example in onnxruntime/test/testdata/custom_op_library/custom_op_library.cc |
CKernelInfo | This struct owns the OrtKernInfo* pointer when a copy is made. For convenient wrapping of OrtKernelInfo* passed to kernel constructor and query attributes, warp the pointer with Ort::Unowned<KernelInfo> instance so it does not destroy the pointer the kernel does not own |
CLogger | This class represents an ONNX Runtime logger that can be used to log information with an associated severity level and source code location (file path, line number, function name) |
CLoraAdapter | LoraAdapter holds a set of Lora Parameters loaded from a single file |
CMapTypeInfo | Wrapper around OrtMapTypeInfo |
CMemoryAllocation | Represents native memory allocation coming from one of the OrtAllocators registered with OnnxRuntime. Use it to wrap an allocation made by an allocator so it can be automatically released when no longer needed |
CMemoryInfo | Wrapper around OrtMemoryInfo |
CModelMetadata | Wrapper around OrtModelMetadata |
COp | Create and own custom defined operation |
COpAttr | This struct provides life time management for custom op attribute |
CRunOptions | RunOptions |
CSequenceTypeInfo | Wrapper around OrtSequenceTypeInfo |
CSession | Wrapper around OrtSession |
CSessionOptions | Wrapper around OrtSessionOptions |
▼CShapeInferContext | Provide access to per-node attributes and input shapes, so one could compute and set output shapes |
CSymbolicInteger | |
CStatus | The Status that holds ownership of OrtStatus received from C API Use it to safely destroy OrtStatus* returned from the C API. Use appropriate constructors to construct an instance of a Status object from exceptions |
CTensorTypeAndShapeInfo | Wrapper around OrtTensorTypeAndShapeInfo |
CThreadingOptions | The ThreadingOptions |
CTrainingSession | Trainer class that provides training, evaluation and optimizer methods for training an ONNX models |
CTypeInfo | Type information that may contain either TensorTypeAndShapeInfo or the information about contained sequence or map depending on the ONNXType |
CValue | Wrapper around OrtValue |
COrtAllocator | Memory allocation interface |
COrtApi | The C API |
COrtApiBase | The helper interface to get the right version of OrtApi |
COrtCUDAProviderOptions | CUDA Provider Options |
COrtCustomHandleType | |
COrtCustomOp | |
COrtMIGraphXProviderOptions | MIGraphX Provider Options |
COrtOpenVINOProviderOptions | OpenVINO Provider Options |
COrtROCMProviderOptions | ROCM Provider Options |
COrtTensorRTProviderOptions | TensorRT Provider Options |
COrtTrainingApi | The Training C API that holds onnxruntime training function pointers |