/// <summary>
/// This is a factory method that creates a disposable instance of FixedBufferOnnxValue
/// on top of a buffer. Internally, it will pin managed buffer and will create
/// an OrtValue containing a tensor that will not own the memory.
/// Such instance of FixedBufferOnnxValue can be used both as input and output in InferenceSession.Run()
/// overload. As compared to CreateFromTensor(), this allows you to pass in buffers with custom data types
/// that are blittable as defined in https://docs.microsoft.com/en-us/dotnet/framework/interop/blittable-and-non-blittable-types
/// I.e. those that have the same binary representation as the original type. This includes all existing types
/// but may also allow using custom types for Float16 and BFloat16 providing they have the same layout and size.
/// The resulting instance must be disposed of to release pinned memory and deallocate native OrtValue
/// See example below.
/// </summary>
/// <typeparam name="T">Blittable data type, compatible with supported types</typeparam>
/// <param name="memoryInfo">memoryInfo. For managed buffers simply use OrtMemoryInfo.DefaultInstance</param>
/// <param name="memory"></param>
/// <param name="elementType">TensorElementType</param>
/// <param name="shape">shape of the tensor to be created</param>
/// <param name="bytesSize">size of the allocation in bytes</param>
/// <returns>a disposable instance of FixedBufferOnnxValue</returns>
/// <example>
/// Here is an example of using a 3rd party library class for processing float16/bfloat16.
/// Currently, to pass tensor data and create a tensor one must copy data to Float16/BFloat16 structures
/// so DenseTensor can recognize it.
///
/// If you are using a library that has a class Half and it is blittable, that is its managed in memory representation
/// matches native one and its size is 16-bits, you can use the following conceptual example
/// to feed/fetch data for inference using Half array. This allows you to avoid copying data from your Half[] to Float16[]
///
/// \code{.cs}
/// unsafe { Debug.Assert(sizeof(ushort) == sizeof(Half)); }
/// Half[] input = new Half[] { 5646, 12345 };
/// var input_shape = new long[] {input.Length};
/// Half[] output = new Half[40]; // Whatever the expected len/shape is must match
/// var output_shape = new long[] {output.Length};
///
/// var memInfo = OrtMemoryInfo.DefaultInstance; // CPU
///
/// using(var fixedBufferInput = FixedBufferOnnxvalue.CreateFromMemory<Half>(memInfo,
/// input, TensorElementType.Float16, input_shape, input.Length * sizeof(ushort))
/// using(var fixedBufferOutput = FixedBufferOnnxvalue.CreateFromMemory<Half>(memInfo,
/// output, TensorElementType.Float16, output_shape, output.Length * sizeof(ushort))
/// {
/// FixedBufferOnnxvalue[] inputValues = new FixedBufferOnnxvalue[]{fixedBufferInput};
/// FixedBufferOnnxvalue[] outputValues = new FixedBufferOnnxvalue[]{fixedBufferOutput};
/// session.Run(inputNames, inputValues, outputNames, outputValues);
/// // Output is now in output[]
/// }
/// \endcode
/// </example>
public static FixedBufferOnnxValue CreateFromMemory <T>(OrtMemoryInfo memoryInfo, Memory <T> memory,
TensorElementType elementType, long[] shape, long bytesSize)
{
if (elementType == TensorElementType.String)
{
throw new ArgumentException("String data type is not supported");
}
var memHandle = memory.Pin();
try
{
IntPtr memPtr;
unsafe
{
memPtr = (IntPtr)memHandle.Pointer;
}
var ortValue = OrtValue.CreateTensorValueWithData(memoryInfo,
elementType,
shape,
memPtr, bytesSize);
return(new FixedBufferOnnxValue(memHandle, ortValue, OnnxValueType.ONNX_TYPE_TENSOR, elementType));
}
catch (Exception)
{
memHandle.Dispose();
throw;
}
}
/// <summary>
/// Bind model output to an OrtValue as Tensor with a given type and shape. An instance of OrtMemoryAllocaiton
/// owns the memory and should be alive for the time of execution.The size of the allocation can not be less than required
/// by the Tensor of the given size.
/// </summary>
/// <param name="name">of the output</param>
/// <param name="elementType">tensor element type</param>
/// <param name="shape">tensor shape</param>
/// <param name="allocation">allocated memory</param>
public void BindOutput(string name, Tensors.TensorElementType elementType, long[] shape, OrtMemoryAllocation allocation)
{
using (var ortValue = OrtValue.CreateTensorValueWithData(allocation.Info,
elementType,
shape,
allocation.Pointer, allocation.Size))
BindInputOrOutput(name, ortValue.Handle, false);
}
/// <summary>
/// Bind external allocation as input or output.
/// The allocation is owned by the user code.
/// </summary>
/// <param name="name">name </param>
/// <param name="allocation">non ort allocated memory</param>
/// <param name="isInput">whether this is an input or output</param>
private void BindExternalAllocation(string name, OrtExternalAllocation allocation, bool isInput)
{
using (var ortValue = OrtValue.CreateTensorValueWithData(allocation.Info,
allocation.ElementType,
allocation.Shape,
allocation.Pointer,
allocation.Size))
BindInputOrOutput(name, ortValue.Handle, isInput);
}