Interface IOnnxModel<T>

Namespace

AiDotNet.Interfaces

Assembly

AiDotNet.dll

Defines the contract for ONNX model wrappers that provide cross-platform model inference.

public interface IOnnxModel<T> : IDisposable

Type Parameters

T

The numeric type used for calculations.

Inherited Members

IDisposable.Dispose()

Remarks

This interface provides a unified way to work with ONNX models in AiDotNet. It supports loading models from files, byte arrays, or URLs, and provides methods for running inference with AiDotNet Tensor types.

For Beginners: ONNX (Open Neural Network Exchange) is a universal format for neural network models. This interface allows you to:

• Load models trained in PyTorch, TensorFlow, or other frameworks
• Run inference using CPU, GPU (CUDA), TensorRT, or DirectML
• Convert between AiDotNet tensors and ONNX tensors automatically

Properties

ExecutionProvider

Gets the execution provider currently being used (CPU, CUDA, TensorRT, DirectML).

string ExecutionProvider { get; }

Property Value

string

IsLoaded

Gets whether the model has been successfully loaded and is ready for inference.

bool IsLoaded { get; }

Property Value

bool

Metadata

Gets the metadata about the loaded ONNX model.

IOnnxModelMetadata Metadata { get; }

Property Value

IOnnxModelMetadata

Methods

Run(Tensor<T>)

Runs inference with a single input tensor.

Tensor<T> Run(Tensor<T> input)

Parameters

input Tensor<T>

The input tensor.

Returns

Tensor<T>

The output tensor from the model.

Run(IReadOnlyDictionary<string, Tensor<T>>)

Runs inference with named inputs.

IReadOnlyDictionary<string, Tensor<T>> Run(IReadOnlyDictionary<string, Tensor<T>> inputs)

Parameters

inputs IReadOnlyDictionary<string, Tensor<T>>

Dictionary mapping input names to tensors.

Returns

IReadOnlyDictionary<string, Tensor<T>>

Dictionary mapping output names to tensors.

RunAsync(Tensor<T>, CancellationToken)

Runs inference asynchronously with a single input tensor.

Task<Tensor<T>> RunAsync(Tensor<T> input, CancellationToken cancellationToken = default)

Parameters

input Tensor<T>

The input tensor.

cancellationToken CancellationToken

Cancellation token.

Returns

Task<Tensor<T>>

The output tensor from the model.

RunAsync(IReadOnlyDictionary<string, Tensor<T>>, CancellationToken)

Runs inference asynchronously with named inputs.

Task<IReadOnlyDictionary<string, Tensor<T>>> RunAsync(IReadOnlyDictionary<string, Tensor<T>> inputs, CancellationToken cancellationToken = default)

Parameters

inputs IReadOnlyDictionary<string, Tensor<T>>

Dictionary mapping input names to tensors.

cancellationToken CancellationToken

Cancellation token.

Returns

Task<IReadOnlyDictionary<string, Tensor<T>>>

Dictionary mapping output names to tensors.

WarmUp()

Warms up the model by running a single inference with dummy data. This helps ensure consistent inference times by initializing lazy resources.

void WarmUp()

WarmUpAsync(CancellationToken)

Warms up the model asynchronously.

Task WarmUpAsync(CancellationToken cancellationToken = default)

Parameters

cancellationToken CancellationToken

Cancellation token.

Returns

Task