AugmentationPipeline<T, TData>
Models & Types · AiDotNet.Augmentation
Represents a pipeline of augmentations that are applied in sequence or composition.
Implements: IAugmentationPolicy<T, TData>
For Beginners
Think of this as a recipe of transformations. You might want to first flip an image, then rotate it, then adjust the colors. This pipeline handles all of that automatically.
How It Works
AugmentationPipeline allows you to compose multiple augmentations together. Each augmentation in the pipeline is applied with its own probability, and the order can be sequential, random, or one-of.
Constructors
| Constructor | Summary |
|---|---|
AugmentationPipeline(string name) |
Creates a new augmentation pipeline with an optional name. |
Properties
| Property | Summary |
|---|---|
AugmentationOrder Order { get; set; } |
Gets or sets the application order for augmentations. |
IList<IAugmentation<T, TData>> Augmentations { get; } |
|
IList<string> AugmentationNames { get; } |
|
int AugmentationCount { get; } |
|
string Name { get; set; } |
Methods
| Method | Summary |
|---|---|
AugmentationPipeline<T, TData> Add(IAugmentation<T, TData> augmentation) |
Adds an augmentation to the pipeline. |
AugmentationPipeline<T, TData> AddRange(IEnumerable<IAugmentation<T, TData>> augmentations) |
Adds multiple augmentations to the pipeline. |
AugmentationPipeline<T, TData> OneOf(IAugmentation<T, TData>[] augmentations) |
Creates a sub-pipeline that applies one-of the specified augmentations. |
AugmentationPipeline<T, TData> Shuffle(IAugmentation<T, TData>[] augmentations) |
Creates a sub-pipeline with random ordering. |
IDictionary<string, object> GetConfiguration() |
|
TData Apply(TData data, AugmentationContext<T> context) |