Tflite Flutter

https://pub.dev/documentation/tflite_flutter/latest
Admin
TensorFlow Lite Flutter plugin provides a flexible and fast solution for accessing TensorFlow Lite...
Tokens:35,158
Snippets:435
Trust Score:9.9
Update:1 month ago
Show doc for...
Tokens:
Raw
### Initialize workspace with Melos

Source: https://pub.dev/documentation/tflite_flutter/latest/index.html

Run these commands to install Melos globally and bootstrap the project workspace.

```bash
dart pub global activate melos # Install or activate melos globally
melos bootstrap # Initialize the workspace and bootstrap the package
```

--------------------------------

### Usage of Interpreter.fromBuffer

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/Interpreter/Interpreter.fromBuffer.html

Example showing how to load a model file as a Uint8List buffer and initialize the interpreter.

```dart
  final buffer = await getBuffer('assets/your_model.tflite');
  final interpreter = Interpreter.fromBuffer(buffer);

  Future<Uint8List> getBuffer(String filePath) async {
      final rawAssetFile = await rootBundle.load(filePath);
      final rawBytes = rawAssetFile.buffer.asUint8List();
      return rawBytes;
  }
```

--------------------------------

### Configure Windows CMakeLists.txt for TFLite

Source: https://pub.dev/documentation/tflite_flutter/latest/index.html

Append these lines to your windows/CMakeLists.txt to install the TFLite Windows dynamic link library.

```cmake
# get tf lite binaries
install(
  FILES ${PROJECT_BUILD_DIR}/../blobs/libtensorflowlite_c-win.dll 
  DESTINATION ${INSTALL_BUNDLE_DATA_DIR}/../blobs/
)
```

--------------------------------

### int get address

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/Interpreter/address.html

Retrieves the memory address of the interpreter instance.

```APIDOC
## int get address

### Description
Returns the memory address of the underlying TFLite interpreter instance.

### Signature
`int get address`

### Returns
- **int**: The memory address of the interpreter.
```

--------------------------------

### int get zeroPoint

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/QuantizationParams/zeroPoint.html

Retrieves the zero-point value for the quantization parameters.

```APIDOC
## int get zeroPoint

### Description
Returns the zero-point value associated with the quantization parameters. This value is used to map quantized values back to real-world values.

### Signature
`int get zeroPoint`

### Returns
- **int** - The zero-point value.
```

--------------------------------

### Configure Linux CMakeLists.txt for TFLite

Source: https://pub.dev/documentation/tflite_flutter/latest/index.html

Append these lines to your linux/CMakeLists.txt to install the TFLite Linux shared object file.

```cmake
# get tf lite binaries
install(
  FILES ${PROJECT_BUILD_DIR}/../blobs/libtensorflowlite_c-linux.so
  DESTINATION ${INSTALL_BUNDLE_DATA_DIR}/../blobs/
)
```

--------------------------------

### String get name

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/Tensor/name.html

Retrieves the name of the tensor element.

```APIDOC
## String get name

### Description
Returns the name of the tensor element as a String.

### Signature
`String get name`

### Returns
- **String** - The name of the tensor.
```

--------------------------------

### Get TfLiteVersion

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/version.html

Retrieves the TensorFlow Lite version as a Dart string.

```dart
String get version => tfliteBinding.TfLiteVersion().cast<Utf8>().toDartString();
```

--------------------------------

### Usage of Interpreter.fromFile

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/Interpreter/Interpreter.fromFile.html

Demonstrates how to load a model file from assets and initialize an Interpreter. The getFile helper function is required to copy the asset to the device's temporary directory.

```dart
final dataFile = await getFile('assets/your_model.tflite');
final interpreter = Interpreter.fromFile(dataFile);

Future<File> getFile(String fileName) async {
  final appDir = await getTemporaryDirectory();
  final appPath = appDir.path;
  final fileOnDevice = File('$appPath/$fileName');
  final rawAssetFile = await rootBundle.load(fileName);
  final rawBytes = rawAssetFile.buffer.asUint8List();
  await fileOnDevice.writeAsBytes(rawBytes, flush: true);
  return fileOnDevice;
}
```

--------------------------------

### TensorType get type

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/Tensor/type.html

Retrieves the data type of the tensor element.

```APIDOC
## TensorType get type

### Description
Returns the data type of the tensor element as a `TensorType` object.

### Signature
`TensorType get type`

### Returns
- **TensorType** - The data type of the tensor element.
```

--------------------------------

### CoreMlDelegate({CoreMlDelegateOptions? options})

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/CoreMlDelegate/CoreMlDelegate.html

Creates a new instance of the CoreMlDelegate. If no options are provided, it initializes with default CoreMlDelegateOptions.

```APIDOC
## CoreMlDelegate

### Description
Creates a new CoreMlDelegate instance to enable Core ML acceleration for TFLite models on iOS.

### Constructor
`CoreMlDelegate({CoreMlDelegateOptions? options})`

### Parameters
- **options** (CoreMlDelegateOptions) - Optional - Configuration options for the Core ML delegate. If null, default options are used.
```

--------------------------------

### GpuDelegateV2 Constructor

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/GpuDelegateV2-class.html

Creates an instance of GpuDelegateV2. If no options are provided, it uses the default options.

```APIDOC
## GpuDelegateV2 Constructor

### Description
Creates GpuDelegateV2 using the specified `options`. Uses `GpuDelegateOptionsV2.defaultOptions()` if options are not specified.

### Signature
```dart
GpuDelegateV2({GpuDelegateOptionsV2? options})
```
```

--------------------------------

### GpuDelegate()

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/GpuDelegate/GpuDelegate.html

Creates a new GpuDelegate instance, optionally accepting GpuDelegateOptions to configure the GPU delegate settings.

```APIDOC
## GpuDelegate()

### Description
Initializes a new GpuDelegate instance for TFLite. If options are provided, they are used to configure the delegate; otherwise, default settings are applied.

### Parameters
- **options** (GpuDelegateOptions) - Optional - Configuration options for the GPU delegate.
```

--------------------------------

### Get Interpreter address

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/Interpreter/address.html

Returns the memory address of the interpreter instance.

```dart
int get address => _interpreter.address;
```

--------------------------------

### Implementation of fromAsset

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/Interpreter/fromAsset.html

The internal implementation of the fromAsset method, which retrieves the asset as a buffer and initializes the interpreter.

```dart
static Future<Interpreter> fromAsset(String assetName,
    {InterpreterOptions? options}) async {
  Uint8List buffer = await _getBuffer(assetName);
  return Interpreter.fromBuffer(buffer, options: options);
}
```

--------------------------------

### Get TfliteFlutterPlatform instance

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter_platform_interface/TfliteFlutterPlatform/instance.html

Retrieves the default instance of TfliteFlutterPlatform, which defaults to MethodChannelTfliteFlutter.

```dart
static TfliteFlutterPlatform get instance => _instance;
```

--------------------------------

### List<int> get shape

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/Tensor/shape.html

Retrieves the dimensions of the tensor as a list of integers.

```APIDOC
## List<int> get shape

### Description
Returns the dimensions of the tensor as a `List<int>`. Each integer in the list represents the size of the tensor along that specific dimension.

### Signature
`List<int> get shape`

### Returns
- **List<int>** - A list containing the size of each dimension of the tensor.
```

--------------------------------

### Interpreter.fromFile(File modelFile, {InterpreterOptions? options})

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/Interpreter/Interpreter.fromFile.html

Creates an Interpreter instance from a provided model file. Throws an ArgumentError if the initialization is unsuccessful.

```APIDOC
## Interpreter.fromFile

### Description
Creates an Interpreter from a model file located on the device.

### Signature
`Interpreter.fromFile(File modelFile, {InterpreterOptions? options})`

### Parameters
- **modelFile** (File) - Required - The model file to be loaded.
- **options** (InterpreterOptions) - Optional - Configuration options for the interpreter.

### Example
```dart
final dataFile = await getFile('assets/your_model.tflite');
final interpreter = Interpreter.fromFile(dataFile);
```
```

--------------------------------

### double get scale

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/QuantizationParams/scale.html

Retrieves the scale value associated with the quantization parameters.

```APIDOC
## double get scale

### Description
Returns the scale value used for quantization. This property is part of the QuantizationParams class.

### Signature
`double get scale`

### Returns
- **double** - The scale factor used in the quantization process.
```

--------------------------------

### Implementation of Interpreter.fromFile

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/Interpreter/Interpreter.fromFile.html

The internal implementation of the factory constructor, showing how it creates a Model from the file path and manages the model lifecycle.

```dart
factory Interpreter.fromFile(File modelFile, {InterpreterOptions? options}) {
  final model = Model.fromFile(modelFile.path);
  final interpreter = Interpreter._create(model, options: options);
  model.delete();
  return interpreter;
}
```

--------------------------------

### GpuDelegateOptions Constructor

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/GpuDelegateOptions-class.html

Initializes GpuDelegateOptions with customizable parameters for Metal GPU delegate.

```APIDOC
## GpuDelegateOptions

### Description
Constructs a GpuDelegateOptions object with optional parameters to configure the Metal GPU delegate.

### Parameters
* **allowPrecisionLoss** (bool) - Optional - If true, allows precision loss for potential performance gains.
* **waitType** (int) - Optional - Specifies the wait type for the delegate, defaulting to `TFLGpuDelegateWaitType.TFLGpuDelegateWaitTypePassive`.
* **enableQuantization** (bool) - Optional - If true, enables quantization support, defaulting to true.
```

--------------------------------

### Pointer<TfLiteDelegate> get base

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/Delegate/base.html

Retrieves the native pointer to the TfLiteDelegate instance.

```APIDOC
## Property: base

### Description
Returns the native `Pointer<TfLiteDelegate>` associated with the delegate instance. This is used to interface with the underlying TensorFlow Lite C library.

### Signature
`Pointer<TfLiteDelegate> get base`
```

--------------------------------

### XNNPackDelegateOptions

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter

Options for configuring the XNNPack delegate.

```APIDOC
## Class: XNNPackDelegateOptions

### Description
XNNPackDelegate Options.

### Properties
- `numThreads` (int) - Number of threads to use for XNNPack.
```

--------------------------------

### Uint8List get data

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/Tensor/data.html

Retrieves the underlying data buffer of the tensor as an unmodifiable Uint8List.

```APIDOC
## Uint8List get data

### Description
Returns the underlying data buffer of the tensor as a `Uint8List`. The returned list is an unmodifiable view of the buffer.

### Returns
- **Uint8List** - The tensor data as a byte list.
```

--------------------------------

### XNNPackDelegateOptions Constructor

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/XNNPackDelegateOptions-class.html

Initializes XNNPackDelegateOptions with a specified number of threads.

```APIDOC
## XNNPackDelegateOptions

### Description
Initializes XNNPackDelegateOptions with an optional number of threads.

### Parameters
#### Named Parameters
- **numThreads** (int) - Optional - The number of threads to use for XNNPack. Defaults to 1.
```

--------------------------------

### Get Tensor data property

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/Tensor/data.html

Retrieves the underlying data buffer as an unmodifiable Uint8List.

```dart
Uint8List get data {
  final data = cast<Uint8>(tfliteBinding.TfLiteTensorData(_tensor));
  return data
      .asTypedList(tfliteBinding.TfLiteTensorByteSize(_tensor))
      .asUnmodifiableView();
}
```

--------------------------------

### Interpreter.fromAsset(String assetName, {InterpreterOptions? options})

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/Interpreter/fromAsset.html

Creates an Interpreter instance by loading a .tflite model file from the specified asset path.

```APIDOC
## Interpreter.fromAsset

### Description
Creates an interpreter from a provided asset name. The .tflite file must be located in the project's assets folder.

### Signature
`Future<Interpreter> fromAsset(String assetName, {InterpreterOptions? options})`

### Parameters
- **assetName** (String) - Required - The path to the .tflite model file within the assets folder.
- **options** (InterpreterOptions) - Optional - Configuration options for the interpreter.

### Example
```dart
final interpreter = await tfl.Interpreter.fromAsset('assets/your_model.tflite');
```
```

--------------------------------

### List<int> get shape

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/ListShape/shape.html

Retrieves the shape of the list by traversing nested elements.

```APIDOC
## List<int> get shape

### Description
Returns the shape of the list as a List of integers. If the list is empty, it returns an empty list.

### Signature
`List<int> get shape`

### Returns
- **List<int>** - A list representing the dimensions of the nested structure.
```

--------------------------------

### bool get isDeleted

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/Interpreter/isDeleted.html

Returns a boolean value indicating if the interpreter instance is deleted.

```APIDOC
## bool get isDeleted

### Description
Returns true if the interpreter instance has been deleted, otherwise returns false.

### Signature
`bool get isDeleted`

### Returns
- **bool** - The deletion status of the interpreter.
```

--------------------------------

### GpuDelegateOptionsV2 Constructor

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/GpuDelegateOptionsV2-class.html

Creates GpuDelegateOptionsV2 with specified parameters. This constructor allows customization of GPU delegate behavior, including precision, inference preference, and priority.

```APIDOC
## GpuDelegateOptionsV2

### Description
Creates GpuDelegateOptionsV2 with specified parameters.

### Parameters
- **isPrecisionLossAllowed** (bool) - Optional - Defaults to `false`. Indicates if precision loss is allowed.
- **inferencePreference** (int) - Optional - Defaults to `TfLiteGpuInferenceUsage.TFLITE_GPU_INFERENCE_PREFERENCE_FAST_SINGLE_ANSWER`. Sets the inference preference.
- **inferencePriority1** (int) - Optional - Defaults to `TfLiteGpuInferencePriority.TFLITE_GPU_INFERENCE_PRIORITY_MAX_PRECISION`. Sets the first priority for inference.
- **inferencePriority2** (int) - Optional - Defaults to `TfLiteGpuInferencePriority.TFLITE_GPU_INFERENCE_PRIORITY_AUTO`. Sets the second priority for inference.
- **inferencePriority3** (int) - Optional - Defaults to `TfLiteGpuInferencePriority.TFLITE_GPU_INFERENCE_PRIORITY_AUTO`. Sets the third priority for inference.
- **experimentalFlags** (List<int>) - Optional - Defaults to `[TfLiteGpuExperimentalFlags.TFLITE_GPU_EXPERIMENTAL_FLAGS_ENABLE_QUANT]`. A list of experimental flags to enable.
- **maxDelegatePartitions** (int) - Optional - Defaults to `1`. The maximum number of delegate partitions.
```

--------------------------------

### ByteConversionUtils()

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/ByteConversionUtils/ByteConversionUtils.html

Constructor for the ByteConversionUtils class. Initializes a new instance of the utility class.

```APIDOC
## ByteConversionUtils()

### Description
Constructor for the ByteConversionUtils class. Initializes a new instance of the utility class.

### Method
Constructor

### Endpoint
N/A

### Parameters
None

### Response
N/A
```

--------------------------------

### int get computeNumElements

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/ListShape/computeNumElements.html

Retrieves the total number of elements in the list based on its shape dimensions.

```APIDOC
## int get computeNumElements

### Description
Calculates and returns the total number of elements in the list by multiplying all dimensions defined in the shape.

### Returns
- **int** - The total count of elements.
```

--------------------------------

### int get lastNativeInferenceDurationMicroSeconds

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/Interpreter/lastNativeInferenceDurationMicroSeconds.html

Retrieves the duration of the last native inference performed by the interpreter in microseconds.

```APIDOC
## int get lastNativeInferenceDurationMicroSeconds

### Description
Returns the duration of the last native inference execution in microseconds.

### Return Value
- **int** - The duration in microseconds.
```

--------------------------------

### CoreMlDelegateOptions()

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/CoreMlDelegateOptions/CoreMlDelegateOptions.html

Creates a new instance of CoreMlDelegateOptions with specified configuration parameters for the CoreML delegate.

```APIDOC
## CoreMlDelegateOptions()

### Description
Factory constructor to initialize CoreML delegate options for TFLite.

### Parameters
- **enabledDevices** (int) - Optional - Specifies which devices are enabled for the delegate. Defaults to TfLiteCoreMlDelegateDevicesWithNeuralEngine.
- **coremlVersion** (int) - Optional - Sets the CoreML version. Defaults to 0.
- **maxDelegatedPartitions** (int) - Optional - Sets the maximum number of delegated partitions. Defaults to 0.
- **minNodesPerPartition** (int) - Optional - Sets the minimum number of nodes per partition. Defaults to 2.
```

--------------------------------

### GpuDelegateOptions()

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/GpuDelegateOptions/GpuDelegateOptions.html

Creates a new instance of GpuDelegateOptions with specified configuration parameters for the GPU delegate.

```APIDOC
## GpuDelegateOptions()

### Description
Creates a new GpuDelegateOptions object to configure GPU acceleration for TFLite models.

### Parameters
- **allowPrecisionLoss** (bool) - Optional - Whether to allow precision loss (default: false).
- **waitType** (int) - Optional - The wait type for the GPU delegate (default: TFLGpuDelegateWaitType.TFLGpuDelegateWaitTypePassive).
- **enableQuantization** (bool) - Optional - Whether to enable quantization (default: true).
```

--------------------------------

### CoreMlDelegateOptions Constructor

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/CoreMlDelegateOptions-class.html

Initializes a CoreMlDelegateOptions object with specified parameters.

```APIDOC
## CoreMlDelegateOptions

### Description
Initializes a CoreMlDelegateOptions object with specified parameters.

### Parameters
- **enabledDevices** (int) - Optional - Specifies the enabled devices for the Core ML delegate. Defaults to `TfLiteCoreMlDelegateEnabledDevices.TfLiteCoreMlDelegateDevicesWithNeuralEngine`.
- **coremlVersion** (int) - Optional - Specifies the Core ML version. Defaults to 0.
- **maxDelegatedPartitions** (int) - Optional - Specifies the maximum number of delegated partitions. Defaults to 0.
- **minNodesPerPartition** (int) - Optional - Specifies the minimum number of nodes per partition. Defaults to 2.
```

--------------------------------

### bool get isAllocated

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/Interpreter/isAllocated.html

Returns true if the interpreter has successfully allocated memory for the model, otherwise returns false.

```APIDOC
## bool get isAllocated

### Description
Returns a boolean value indicating the allocation status of the interpreter.

### Signature
`bool get isAllocated`

### Returns
- **bool**: True if the interpreter is allocated, false otherwise.
```

--------------------------------

### InterpreterOptions() Constructor

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/InterpreterOptions-class.html

Creates a new InterpreterOptions instance.

```APIDOC
## InterpreterOptions()

### Description
Creates a new options instance.

### Constructor
`InterpreterOptions()`
```

--------------------------------

### GpuDelegateOptions

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/tflite_flutter-library.html

Options for configuring the GpuDelegate.

```APIDOC
## Class: GpuDelegateOptions

### Description
Metal Delegate options.

### Properties
- `allowPrecisionLoss` (bool)
  Whether to allow precision loss for performance.
- `waitForGrBe` (bool)
  Whether to wait for GPU to finish rendering.
```

--------------------------------

### Create Interpreter from Asset

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/Interpreter/fromAsset.html

Basic usage of the fromAsset method to initialize an interpreter. Ensure the .tflite file is located in the project's assets folder.

```dart
final interpreter = await tfl.Interpreter.fromAsset('assets/your_model.tflite');
```

--------------------------------

### Get tensor name implementation

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/Tensor/name.html

Retrieves the name of the tensor element by calling the TfLiteTensorName binding and converting the result to a Dart string.

```dart
String get name =>
    tfliteBinding.TfLiteTensorName(_tensor).cast<Utf8>().toDartString();
```

--------------------------------

### Interpreter.fromBuffer(Uint8List buffer, {InterpreterOptions? options})

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/Interpreter/Interpreter.fromBuffer.html

Creates an instance of the Interpreter from a provided model buffer. Throws an ArgumentError if the initialization fails.

```APIDOC
## Interpreter.fromBuffer

### Description
Creates an interpreter from a `Uint8List` buffer containing the model data. This is useful for loading models directly from memory.

### Signature
`Interpreter.fromBuffer(Uint8List buffer, {InterpreterOptions? options})`

### Parameters
- **buffer** (Uint8List) - Required - The model data as a byte array.
- **options** (InterpreterOptions) - Optional - Configuration options for the interpreter.

### Example
```dart
final buffer = await getBuffer('assets/your_model.tflite');
final interpreter = Interpreter.fromBuffer(buffer);
```
```

--------------------------------

### CoreMlDelegateOptions

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/tflite_flutter-library.html

Options for configuring the CoreMlDelegate.

```APIDOC
## Class: CoreMlDelegateOptions

### Description
CoreMlDelegate Options.

### Properties
- `coremlVersion` (int?)
  The Core ML version to use.
- `enabledDevices` (int?)
  The enabled devices for Core ML.
```

--------------------------------

### Tensor Class Properties

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/Tensor-class.html

Access properties of a Tensor object to get information about its data, name, shape, type, and quantization parameters.

```APIDOC
## Properties

data ↔ Uint8List
     Underlying data buffer as bytes. 
getter/setter pair 

hashCode → int 
     The hash code for this object. 
no setterinherited 

name → String 
     Name of the tensor element. 
no setter 

params → QuantizationParams 
     Quantization Params associated with the model, `only Android`
no setter 

runtimeType → Type 
     A representation of the runtime type of the object. 
no setterinherited 

shape → List<int> 
     Dimensions of the tensor. 
no setter 

type → TensorType 
     Data type of the tensor element. 
no setter
```

--------------------------------

### Implementation of shape property

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/Tensor/shape.html

This snippet shows the implementation of the shape getter, which uses the tfliteBinding to retrieve the number of dimensions and then iterates to get each dimension size.

```dart
List<int> get shape => List.generate(
    tfliteBinding.TfLiteTensorNumDims(_tensor),
    (i) => tfliteBinding.TfLiteTensorDim(_tensor, i));
```

--------------------------------

### XNNPackDelegateOptions Methods

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/XNNPackDelegateOptions-class.html

Methods available for XNNPackDelegateOptions.

```APIDOC
## Methods

### delete
- **Returns**: `void`
- **Description**: Releases the resources associated with the XNNPack delegate options.
```

```APIDOC
### noSuchMethod
- **Parameters**:
  - **invocation** (Invocation) - The invocation details.
- **Returns**: `dynamic`
- **Description**: Invoked when a nonexistent method or property is accessed. Inherited.
```

```APIDOC
### toString
- **Returns**: `String`
- **Description**: A string representation of this object. Inherited.
```

--------------------------------

### XNNPackDelegate Constructor

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/XNNPackDelegate-class.html

Constructs an instance of XNNPackDelegate. Optionally accepts XNNPackDelegateOptions for configuration.

```APIDOC
## Constructors

XNNPackDelegate({XNNPackDelegateOptions? options})

### Description
Constructs an instance of XNNPackDelegate. Optionally accepts XNNPackDelegateOptions for configuration.

### Parameters
* **options** (XNNPackDelegateOptions?) - Optional - Configuration options for the XNNPack delegate.
```

--------------------------------

### GpuDelegateOptions factory constructor implementation

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/GpuDelegateOptions/GpuDelegateOptions.html

This factory constructor initializes TFLGpuDelegateOptions using calloc and sets the provided configuration parameters.

```dart
factory GpuDelegateOptions({
  bool allowPrecisionLoss = false,
  int waitType = TFLGpuDelegateWaitType.TFLGpuDelegateWaitTypePassive,
  bool enableQuantization = true,
}) {
  final options = calloc<TFLGpuDelegateOptions>();

  options.ref
    ..allow_precision_loss = allowPrecisionLoss
    ..wait_type = waitType
    ..enable_quantization = enableQuantization;

  return GpuDelegateOptions._(options);
}
```

--------------------------------

### GpuDelegate Methods

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/GpuDelegate-class.html

Provides methods for managing the delegate instance and object representation.

```APIDOC
## Methods

### delete
* `void delete()`
  Destroys delegate instance. Overrides base method.

### noSuchMethod
* `dynamic noSuchMethod(Invocation invocation)`
  Invoked when a nonexistent method or property is accessed. Inherited.

### toString
* `String toString()`
  A string representation of this object. Inherited.
```

--------------------------------

### GpuDelegate factory constructor implementation

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/GpuDelegate/GpuDelegate.html

The factory constructor initializes a GpuDelegate, using default settings if options are null or the provided GpuDelegateOptions if specified.

```dart
factory GpuDelegate({GpuDelegateOptions? options}) {
  if (options == null) {
    return GpuDelegate._(
      tfliteBinding.TFLGpuDelegateCreate(nullptr),
    );
  }
  return GpuDelegate._(tfliteBinding.TFLGpuDelegateCreate(options.base));
}
```

--------------------------------

### void runInference(List<Object> inputs)

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/Interpreter/runInference.html

Executes the inference process for the loaded model using the provided input data.

```APIDOC
## void runInference(List<Object> inputs)

### Description
Runs inference on the loaded model. This method handles input tensor validation, dynamic resizing of input tensors if the input shape differs, memory allocation, and the execution of the native inference call.

### Parameters
- **inputs** (List<Object>) - Required - A list of input objects to be processed by the model.

### Throws
- **ArgumentError** - Thrown if the provided inputs list is empty.
```

--------------------------------

### Create Interpreter from Asset

Source: https://pub.dev/documentation/tflite_flutter/latest/index.html

Create an interpreter instance from a model file located in the assets directory.

```dart
final interpreter = await Interpreter.fromAsset('assets/your_model.tflite');
```

--------------------------------

### void run(Object input, Object output)

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/Interpreter/run.html

Executes the TFLite model for a single input and stores the result in the provided output object.

```APIDOC
## void run(Object input, Object output)

### Description
Runs the interpreter for a single input and output pair. This method internally maps the output to index 0 and calls the multi-input execution logic.

### Parameters
- **input** (Object) - The input data for the model.
- **output** (Object) - The object where the inference result will be stored.

### Implementation
```dart
void run(Object input, Object output) {
  var map = <int, Object>{};
  map[0] = output;
  runForMultipleInputs([input], map);
}
```
```

--------------------------------

### Implementation of run method

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/Interpreter/run.html

This implementation maps the provided output to index 0 and delegates the execution to runForMultipleInputs.

```dart
void run(Object input, Object output) {
  var map = <int, Object>{};
  map[0] = output;
  runForMultipleInputs([input], map);
}
```

--------------------------------

### GpuDelegateV2 factory constructor implementation

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/GpuDelegateV2/GpuDelegateV2.html

The factory constructor creates a GpuDelegateV2 instance. It uses the provided GpuDelegateOptionsV2 or defaults to a null pointer if no options are specified.

```dart
factory GpuDelegateV2({GpuDelegateOptionsV2? options}) {
  if (options == null) {
    return GpuDelegateV2._(
      tfliteBindingGpu.TfLiteGpuDelegateV2Create(nullptr),
    );
  }
  return GpuDelegateV2._(
      tfliteBindingGpu.TfLiteGpuDelegateV2Create(options.base));
}
```

--------------------------------

### Implementation of Interpreter.fromBuffer

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/Interpreter/Interpreter.fromBuffer.html

The internal implementation of the factory constructor, which creates a model from the buffer and then initializes the interpreter.

```dart
factory Interpreter.fromBuffer(Uint8List buffer,
    {InterpreterOptions? options}) {
  final model = Model.fromBuffer(buffer);
  final interpreter = Interpreter._create(model, options: options);
  model.delete();
  return interpreter;
}
```

--------------------------------

### runForMultipleInputs(List<Object> inputs, Map<int, Object> outputs)

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/Interpreter/runForMultipleInputs.html

Executes inference on the interpreter using a list of input objects and populates the provided map of output objects.

```APIDOC
## runForMultipleInputs(List<Object> inputs, Map<int, Object> outputs)

### Description
Runs inference for multiple inputs and maps the resulting output tensors to the provided output map.

### Parameters
- **inputs** (List<Object>) - A list of input data objects to be processed by the model.
- **outputs** (Map<int, Object>) - A map where keys are output indices and values are the buffers to store the inference results.

### Errors
- Throws an `ArgumentError` if the `outputs` map is empty.
```

--------------------------------

### Delegate Class API

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/Delegate-class.html

This section details the constructors, properties, and methods available for the abstract Delegate class.

```APIDOC
## Delegate Class

### Description
Abstract base class for delegates in tflite_flutter.

### Implementers
- CoreMlDelegate
- GpuDelegate
- GpuDelegateV2
- XNNPackDelegate

## Constructors

### Delegate()

Default constructor for the Delegate class.

## Properties

### base → Pointer<TfLiteDelegate>

#### Description
Get pointer to TfLiteDelegate.

#### Setter
no setter

### hashCode → int

#### Description
The hash code for this object.

#### Setter
no setter

#### Inherited
yes

### runtimeType → Type

#### Description
A representation of the runtime type of the object.

#### Setter
no setter

#### Inherited
yes

## Methods

### delete() → void

#### Description
Destroys delegate instance.

### noSuchMethod(Invocation invocation) → dynamic

#### Description
Invoked when a nonexistent method or property is accessed.

#### Inherited
yes

### toString() → String

#### Description
A string representation of this object.

#### Inherited
yes

## Operators

### operator ==(Object other) → bool

#### Description
The equality operator.

#### Inherited
yes
```

--------------------------------

### Interpreter.run

Source: https://pub.dev/documentation/tflite_flutter/latest/index.html

Performs inference for a model with a single input and output.

```APIDOC
## Interpreter.run

### Description
Executes inference on the model using a single input and output tensor.

### Signature
`void run(Object input, Object output)`

### Parameters
- **input** (Object) - Required - The input data for the model.
- **output** (Object) - Required - The container for the model output.
```

--------------------------------

### CoreMlDelegate Constructors

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/CoreMlDelegate-class.html

Constructs a CoreMlDelegate instance. Options can be provided to configure the delegate.

```APIDOC
## CoreMlDelegate

### Description
Instantiates a CoreMlDelegate. Optionally, you can pass a `CoreMlDelegateOptions` object to configure the delegate's behavior, such as specifying the coreml version or hardware acceleration.

### Constructors
#### CoreMlDelegate

```dart
CoreMlDelegate({CoreMlDelegateOptions? options})
```

- **options** (CoreMlDelegateOptions?) - Optional configuration options for the Core ML delegate.
```

--------------------------------

### TfliteFlutterPlatform()

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter_platform_interface/TfliteFlutterPlatform-class.html

Constructs a new instance of the TfliteFlutterPlatform.

```APIDOC
## Constructors 

TfliteFlutterPlatform() 
     Constructs a TfliteFlutterPlatform. 

```

--------------------------------

### XNNPackDelegate

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter

XNNPack delegate for leveraging XNNPACK library for optimized inference.

```APIDOC
## Class: XNNPackDelegate

### Description
XNNPack Delegate.

### Methods
- `XNNPackDelegate(options: XNNPackDelegateOptions?)`
```

--------------------------------

### runForMultipleInputs(List<Object> inputs, Map<int, Object> outputs)

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/IsolateInterpreter/runForMultipleInputs.html

Executes the TensorFlow model using the provided list of inputs and populates the provided map of output tensors.

```APIDOC
## runForMultipleInputs(List<Object> inputs, Map<int, Object> outputs)

### Description
Runs the TensorFlow model for multiple inputs and maps the resulting output tensors into the provided output map.

### Parameters
- **inputs** (List<Object>) - A list of input objects to be processed by the model.
- **outputs** (Map<int, Object>) - A map where keys are output indices and values are the objects to store the output data.

### Returns
- **Future<void>** - A future that completes when the model execution and output copying are finished.
```

--------------------------------

### Interpreter Constructors

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/Interpreter-class.html

Constructors for creating an Interpreter instance.

```APIDOC
## Interpreter.fromAddress

### Description
Creates interpreter from an address.

### Factory
`Interpreter.fromAddress(int address, {bool allocated = false, bool deleted = false})`
```

```APIDOC
## Interpreter.fromBuffer

### Description
Creates interpreter from a `buffer`.

### Factory
`Interpreter.fromBuffer(Uint8List buffer, {InterpreterOptions? options})`
```

```APIDOC
## Interpreter.fromFile

### Description
Creates Interpreter from a model file.

### Factory
`Interpreter.fromFile(File modelFile, {InterpreterOptions? options})`
```

```APIDOC
## Interpreter.fromAsset

### Description
Creates interpreter from an `assetName`.

### Static Method
`fromAsset(String assetName, {InterpreterOptions? options}) → Future<Interpreter>`
```

--------------------------------

### Implementation of getPlatformVersion

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter_method_channel/MethodChannelTfliteFlutter/getPlatformVersion.html

This method uses the methodChannel to invoke the 'getPlatformVersion' call and returns the resulting string asynchronously.

```dart
@override
Future<String?> getPlatformVersion() async {
  final version =
      await methodChannel.invokeMethod<String>('getPlatformVersion');
  return version;
}
```

--------------------------------

### GpuDelegateOptionsV2.base

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/GpuDelegateOptionsV2/base.html

Accesses the native pointer for the GPU delegate options.

```APIDOC
## Property: base

### Description
Returns the underlying native pointer to the `TfLiteGpuDelegateOptionsV2` structure.

### Signature
`Pointer<TfLiteGpuDelegateOptionsV2> get base`

### Returns
- **Pointer<TfLiteGpuDelegateOptionsV2>** - A pointer to the native GPU delegate options configuration.
```

--------------------------------

### GpuDelegateOptions Methods

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/GpuDelegateOptions-class.html

Provides methods for managing the GpuDelegateOptions object.

```APIDOC
## GpuDelegateOptions Methods

### delete
* **Returns**: `void`
* **Description**: Releases the resources associated with the GpuDelegateOptions.

### noSuchMethod
* **Parameters**:
    * **invocation** (Invocation) - The invocation to handle.
* **Returns**: `dynamic`
* **Description**: Invoked when a nonexistent method or property is accessed. Inherited from `Object`.

### toString
* **Returns**: `String`
* **Description**: A string representation of this object. Inherited from `Object`.
```

--------------------------------

### QuantizationParams constructor implementation

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/QuantizationParams/QuantizationParams.html

The constructor initializes the QuantizationParams object with the provided scale and zero point values.

```dart
QuantizationParams(this._scale, this._zeroPoint);
```

--------------------------------

### GpuDelegateOptionsV2

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/tflite_flutter-library.html

Options for configuring the GpuDelegate on Android.

```APIDOC
## Class: GpuDelegateOptionsV2

### Description
GPU delegate options for Android.

### Properties
- `inferencePriority1` (TfLiteGpuInferencePriority)
  Inference priority 1.
- `inferencePriority2` (TfLiteGpuInferencePriority)
  Inference priority 2.
- `inferencePriority3` (TfLiteGpuInferencePriority)
  Inference priority 3.
- `precisionLossAllowed` (bool)
  Whether precision loss is allowed.
- `experimentalF16PrecisionForAndroid` (bool)
  Whether to use experimental FP16 precision for Android.
```

--------------------------------

### GpuDelegate Constructor

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/GpuDelegate-class.html

Constructs a GpuDelegate instance. Optionally, GpuDelegateOptions can be provided to configure the delegate.

```APIDOC
## GpuDelegate

### Description
Constructs a GpuDelegate instance. Optionally, GpuDelegateOptions can be provided to configure the delegate.

### Constructors

* `GpuDelegate({GpuDelegateOptions? options})`
```

--------------------------------

### GpuDelegateOptionsV2 Factory Implementation

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/GpuDelegateOptionsV2/GpuDelegateOptionsV2.html

The factory constructor initializes the GpuDelegateOptionsV2 object by allocating memory for TfLiteGpuDelegateOptionsV2 and mapping the provided Dart parameters to the underlying C structure.

```dart
factory GpuDelegateOptionsV2({
  bool isPrecisionLossAllowed = false,
  int inferencePreference = TfLiteGpuInferenceUsage
      .TFLITE_GPU_INFERENCE_PREFERENCE_FAST_SINGLE_ANSWER,
  int inferencePriority1 =
      TfLiteGpuInferencePriority.TFLITE_GPU_INFERENCE_PRIORITY_MAX_PRECISION,
  int inferencePriority2 =
      TfLiteGpuInferencePriority.TFLITE_GPU_INFERENCE_PRIORITY_AUTO,
  int inferencePriority3 =
      TfLiteGpuInferencePriority.TFLITE_GPU_INFERENCE_PRIORITY_AUTO,
  List<int> experimentalFlags = const [
    TfLiteGpuExperimentalFlags.TFLITE_GPU_EXPERIMENTAL_FLAGS_ENABLE_QUANT
  ],
  int maxDelegatePartitions = 1,
}) {
  final options = calloc<TfLiteGpuDelegateOptionsV2>();
  options.ref
    ..is_precision_loss_allowed = isPrecisionLossAllowed ? 1 : 0
    ..inference_preference = inferencePreference
    ..inference_priority1 = inferencePriority1
    ..inference_priority2 = inferencePriority2
    ..inference_priority3 = inferencePriority3
    ..experimental_flags =
        _TfLiteGpuExperimentalFlagsUtil.getBitmask(experimentalFlags)
    ..max_delegated_partitions = maxDelegatePartitions;

  return GpuDelegateOptionsV2._(options);
}
```

--------------------------------

### Interpreter.fromAsset

Source: https://pub.dev/documentation/tflite_flutter/latest/index.html

Creates an interpreter instance from a model file stored in the assets directory.

```APIDOC
## Interpreter.fromAsset

### Description
Creates an interpreter instance from a TFLite model file located in the assets directory.

### Signature
`static Future<Interpreter> fromAsset(String assetName)`

### Parameters
- **assetName** (String) - Required - The path to the .tflite model file in the assets directory.
```

--------------------------------

### CoreMlDelegateOptions

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter

Options for configuring the Core ML delegate on iOS.

```APIDOC
## Class: CoreMlDelegateOptions

### Description
CoreMlDelegate Options.

### Properties
- `coremlVersion` (int) - The Core ML version to use.
- `enabledDeviceTypes` (List<int>) - List of enabled device types.
- `maxDelegatedPartitions` (int) - Maximum number of delegated partitions.
- `minWeightMapSize` (int) - Minimum weight map size.
- `cacheDir` (String) - Cache directory for Core ML models.
- `cacheFile` (String) - Cache file for Core ML models.
- `useCoreML3` (bool) - Whether to use Core ML 3.
- `useMetalDelegate` (bool) - Whether to use the Metal delegate.
```

--------------------------------

### Interpreter.fromAddress Implementation

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/Interpreter/Interpreter.fromAddress.html

The implementation of the factory constructor that initializes an Interpreter from a pointer address, setting the allocated and deleted states.

```dart
factory Interpreter.fromAddress(int address,
    {bool allocated = false, bool deleted = false}) {
  final interpreter = Pointer<TfLiteInterpreter>.fromAddress(address);
  return Interpreter._(interpreter)
    .._deleted = deleted
    .._allocated = allocated;
}
```

--------------------------------

### XNNPackDelegate factory implementation

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/XNNPackDelegate/XNNPackDelegate.html

The factory constructor initializes the XNNPack delegate, using default settings if no options are provided or the base pointer from the provided options.

```dart
factory XNNPackDelegate({XNNPackDelegateOptions? options}) {
  if (options == null) {
    return XNNPackDelegate._(
      tfliteBinding.TfLiteXNNPackDelegateCreate(nullptr),
    );
  }
  return XNNPackDelegate._(
      tfliteBinding.TfLiteXNNPackDelegateCreate(options.base));
}
```

--------------------------------

### useMetalDelegateForIOS(bool useMetal)

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/InterpreterOptions/useMetalDelegateForIOS.html

Sets whether to use the Metal Delegate for iOS inference. When set to true, the GpuDelegate is added to the interpreter options if the platform is iOS.

```APIDOC
## useMetalDelegateForIOS(bool useMetal)

### Description
Enables or disables the Metal Delegate for iOS platforms to leverage GPU acceleration for TFLite models.

### Parameters
- **useMetal** (bool) - Required - Set to true to enable the Metal Delegate on iOS.
```

--------------------------------

### Interpreter.fromAddress(int address, {bool allocated, bool deleted})

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/Interpreter/Interpreter.fromAddress.html

Creates an Interpreter instance from a provided memory address. This is primarily used to share interpreter instances across different isolates.

```APIDOC
## Interpreter.fromAddress

### Description
Creates an Interpreter instance from a raw memory address. This is typically used for passing interpreter instances between isolates.

### Signature
`Interpreter.fromAddress(int address, {bool allocated = false, bool deleted = false})`

### Parameters
- **address** (int) - Required - The memory address of the TfLiteInterpreter pointer.
- **allocated** (bool) - Optional - Indicates if the interpreter is already allocated. Defaults to false.
- **deleted** (bool) - Optional - Indicates if the interpreter has been marked as deleted. Defaults to false.
```

--------------------------------

### CoreMlDelegateOptions Methods

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/CoreMlDelegateOptions-class.html

Methods available for the CoreMlDelegateOptions class.

```APIDOC
## Methods

### delete
- **Returns**: `void`
- **Description**: Releases any resources associated with the Core ML delegate options.
```

```APIDOC
### noSuchMethod
- **Parameters**: 
  - **invocation** (Invocation) - The invocation details.
- **Returns**: `dynamic`
- **Description**: Invoked when a nonexistent method or property is accessed. Inherited from `Object`.
```

```APIDOC
### toString
- **Returns**: `String`
- **Description**: A string representation of this object. Inherited from `Object`.
```

--------------------------------

### GpuDelegate

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/tflite_flutter-library.html

The GpuDelegate enables hardware acceleration for TensorFlow Lite models using the GPU.

```APIDOC
## Class: GpuDelegate

### Description
Metal Delegate for iOS.

### Methods
- `GpuDelegate({GpuDelegateOptions? options})`
  Creates a new GpuDelegate with optional options.
```

--------------------------------

### QuantizationParams(double scale, int zeroPoint)

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/QuantizationParams/QuantizationParams.html

Initializes a new instance of QuantizationParams with the specified scale and zero-point values.

```APIDOC
## QuantizationParams(double scale, int zeroPoint)

### Description
Creates a new instance of the QuantizationParams class, which holds the quantization parameters (scale and zero-point) required for interpreting quantized tensor data.

### Parameters
- **scale** (double) - The scale value for quantization.
- **zeroPoint** (int) - The zero-point value for quantization.
```

--------------------------------

### Future<String?> getPlatformVersion()

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter_method_channel/MethodChannelTfliteFlutter/getPlatformVersion.html

Retrieves the platform version string via the method channel.

```APIDOC
## Future<String?> getPlatformVersion()

### Description
Retrieves the current platform version string by invoking the 'getPlatformVersion' method on the platform channel.

### Signature
`Future<String?> getPlatformVersion()`

### Returns
- **Future<String?>** - A future that completes with the platform version string, or null if the version could not be retrieved.
```

--------------------------------

### Run ffigen code generation

Source: https://pub.dev/documentation/tflite_flutter/latest/index.html

Use this Melos command to trigger the generation of FFI bindings for the package.

```bash
melos run ffigen 
```

--------------------------------

### CoreMlDelegate factory constructor implementation

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/CoreMlDelegate/CoreMlDelegate.html

This factory constructor initializes the CoreMlDelegate, using default options if none are provided.

```dart
factory CoreMlDelegate({CoreMlDelegateOptions? options}) {
  final delegateOptions = options ?? CoreMlDelegateOptions();

  return CoreMlDelegate._(
    tfliteBinding.TfLiteCoreMlDelegateCreate(delegateOptions.base),
  );
}
```

--------------------------------

### InterpreterOptions

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/tflite_flutter-library.html

InterpreterOptions allows configuring the behavior of the Interpreter, such as setting up delegates.

```APIDOC
## Class: InterpreterOptions

### Description
TensorFlowLite interpreter options.

### Properties
- `threads` (int?)
  Number of threads to use for inference.
- `useGpu` (bool?)
  Whether to use the GPU delegate.
- `useNnApi` (bool?)
  Whether to use the NNAPI delegate (Android only).
- `useCoreML` (bool?)
  Whether to use the CoreML delegate (iOS only).
- `delegates` (List<Delegate>?)
  A list of delegates to use for inference.
```

--------------------------------

### TfliteFlutterPlatform.instance

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter_platform_interface/TfliteFlutterPlatform/instance.html

Retrieves the current default instance of TfliteFlutterPlatform.

```APIDOC
## static TfliteFlutterPlatform get instance

### Description
Returns the default instance of TfliteFlutterPlatform to use. By default, this returns an instance of MethodChannelTfliteFlutter.
```

--------------------------------

### Implementation of useMetalDelegateForIOS

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/InterpreterOptions/useMetalDelegateForIOS.html

This snippet demonstrates how to set the useMetalDelegateForIOS property, checking for the iOS platform before adding the GpuDelegate.

```dart
set useMetalDelegateForIOS(bool useMetal) {
  if (Platform.isIOS) {
    addDelegate(GpuDelegate());
  }
}
```

--------------------------------

### GpuDelegateV2

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter

GPU delegate for Android, enabling hardware-accelerated inference on supported devices.

```APIDOC
## Class: GpuDelegateV2

### Description
GPU delegate for Android.

### Methods
- `GpuDelegateV2(options: GpuDelegateOptionsV2?)`
```

--------------------------------

### XNNPackDelegateOptions Properties

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/XNNPackDelegateOptions-class.html

Properties available for XNNPackDelegateOptions.

```APIDOC
## Properties

### base
- **Type**: `Pointer<TfLiteXNNPackDelegateOptions>`
- **Description**: Returns the base pointer to the TfLiteXNNPackDelegateOptions structure. No setter available.
```

```APIDOC
### hashCode
- **Type**: `int`
- **Description**: The hash code for this object. No setter available. Inherited.
```

```APIDOC
### runtimeType
- **Type**: `Type`
- **Description**: A representation of the runtime type of the object. No setter available. Inherited.
```

--------------------------------

### XNNPackDelegateOptions.base

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/XNNPackDelegateOptions/base.html

Retrieves the pointer to the TfLiteXNNPackDelegateOptions structure.

```APIDOC
## Property: base

### Description
Returns the underlying pointer to the `TfLiteXNNPackDelegateOptions` structure used by the XNNPack delegate.

### Signature
`Pointer<TfLiteXNNPackDelegateOptions> get base`

### Returns
- **Pointer<TfLiteXNNPackDelegateOptions>** - A pointer to the delegate options structure.
```

--------------------------------

### TensorType Static Methods

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/TensorType.html

Documentation for the static method `fromValue` which allows creating a TensorType from its integer representation.

```APIDOC
## Static Methods

### fromValue(int tfLiteValue) → TensorType
Creates a `TensorType` from its corresponding integer `tfLiteValue`.
```

--------------------------------

### CoreMlDelegate Methods

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/CoreMlDelegate-class.html

Provides methods for interacting with the Core ML delegate, including destroying the delegate instance.

```APIDOC
## CoreMlDelegate Methods

### delete

```dart
void delete()
```

#### Description
Destroys the Core ML delegate instance and releases associated resources. This method should be called when the delegate is no longer needed to prevent memory leaks.
```

--------------------------------

### QuantizationParams Constructor

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/QuantizationParams-class.html

Constructs a QuantizationParams object with the specified scale and zero-point.

```APIDOC
## QuantizationParams Constructor

### Description
Constructs a QuantizationParams object with the specified scale and zero-point.

### Parameters
#### Path Parameters
- **_scale** (double) - Required - The quantization scale.
- **_zeroPoint** (int) - Required - The quantization zero-point.

### Method
QuantizationParams(_scale: double, _zeroPoint: int)
```

--------------------------------

### Future<void> run(Object input, Object output)

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/IsolateInterpreter/run.html

Executes the TensorFlow model using a single input and a single output object.

```APIDOC
## Future<void> run(Object input, Object output)

### Description
Runs the TensorFlow model for a single input and output. This method internally maps the output to index 0 and calls runForMultipleInputs.

### Parameters
- **input** (Object) - The input data for the model.
- **output** (Object) - The output container where the model result will be stored.

### Returns
- **Future<void>** - A future that completes when the model execution is finished.
```

--------------------------------

### void invoke()

Source: https://pub.dev/documentation/tflite_flutter/latest/tflite_flutter/Interpreter/invoke.html

Runs inference for the loaded graph. This method requires the interpreter to be allocated before it is called.

```APIDOC
## void invoke()

### Description
Runs inference for the loaded graph. This method triggers the execution of the model using the data currently set in the input tensors.

### Usage
```dart
interpreter.invoke();
```

### Requirements
- The interpreter must be allocated before calling this method. If not allocated, it will throw a state error.
```
Tokens:
Raw