# Basic Inference

The quickest way to confirm ONNX Runtime is working in your project.

## What it does

Loads a tiny (185-byte) ONNX model that is **embedded directly in the script**
(no `.onnx` file, no download) and runs it. The model computes, element-wise:

```
output = input * 2 + 1
```

With `input = [1, 2, 3]` the expected `output` is `[3, 5, 7]`.

## How to run

1. Import this sample from the Package Manager (**ONNX Runtime → Samples → Basic
   Inference → Import**).
2. Create an empty GameObject in a scene and add the **Basic Inference Sample**
   component (`BasicInferenceSample`).
3. Enter Play Mode and check the **Console**:

   ```
   [ONNX Runtime] input  = [1, 2, 3]
   [ONNX Runtime] output = [3, 5, 7]  (expected [3, 5, 7])
   [ONNX Runtime] Basic inference succeeded ✅
   ```

## Key API

```csharp
using var session = new InferenceSession(modelBytes);
var inputs = new List<NamedOnnxValue>
{
    NamedOnnxValue.CreateFromTensor("input", new DenseTensor<float>(data, new[] { 1, 3 }))
};
using var results = session.Run(inputs);
float[] output = results.First().AsTensor<float>().ToArray();
```