OnnxModel<T>
Models & Types · AiDotNet.Onnx
A wrapper for ONNX models that provides easy-to-use inference with AiDotNet Tensor types.
Implements: IOnnxModel<T>
For Beginners
Use this class to run pre-trained ONNX models:
How It Works
This class wraps the ONNX Runtime InferenceSession and provides:
- Automatic tensor conversion between AiDotNet and ONNX formats
- Support for multiple execution providers (CPU, CUDA, TensorRT, DirectML)
- Multi-input/multi-output model support
- Warm-up and async inference
Constructors
| Constructor | Summary |
|---|---|
OnnxModel(byte[] modelBytes, OnnxModelOptions options) |
Creates a new OnnxModel from a byte array. |
OnnxModel(string modelPath, OnnxModelOptions options) |
Creates a new OnnxModel from a byte array. |
Properties
| Property | Summary |
|---|---|
IOnnxModelMetadata Metadata { get; } |
|
bool IsLoaded { get; } |
|
string ExecutionProvider { get; } |
Methods
| Method | Summary |
|---|---|
IReadOnlyDictionary<string, Tensor<T>> Run(IReadOnlyDictionary<string, Tensor<T>> inputs) |
|
IReadOnlyDictionary<string, Tensor<T>> Run(IReadOnlyDictionary<string, Tensor<T>> inputs, IEnumerable<string> outputNames) |
Runs inference with specific output names. |
Task WarmUpAsync(CancellationToken cancellationToken) |
|
Task<IReadOnlyDictionary<string, Tensor<T>>> RunAsync(IReadOnlyDictionary<string, Tensor<T>> inputs, CancellationToken cancellationToken) |
|
Task<OnnxModel<T>> CreateAsync(string modelPath, OnnxModelOptions options, IProgress<double> progress, CancellationToken cancellationToken) |
Creates an OnnxModel asynchronously, optionally downloading from a URL. |
Task<Tensor<T>> RunAsync(Tensor<T> input, CancellationToken cancellationToken) |
|
Tensor<T> Run(Tensor<T> input) |
|
Tensor<T> RunWithLongInput(string inputName, long[] tokenIds) |
Runs inference with long integer inputs (useful for token IDs). |
void Dispose() |
Disposes the ONNX session and releases resources. |
void WarmUp() |