LayoutLM<T>
Models & Types · AiDotNet.Document.LayoutAware
LayoutLM (v1) neural network for document understanding with layout-aware pre-training.
Inheritance: NeuralNetworkBase<T> → DocumentNeuralNetworkBase<T> → LayoutLM<T>
Implements: ILayoutDetector<T>
For Beginners
LayoutLM understands documents by learning from both:
- The text content (what the words say)
- The layout structure (where words are positioned on the page)
Unlike LayoutLMv2/v3, this version does NOT use visual features (images), making it lighter but less powerful for visually-rich documents.
Example usage:
How It Works
LayoutLM is the first generation of Microsoft’s layout-aware document understanding models. It combines text embeddings with 2D position embeddings to jointly model text and layout.
Reference: “LayoutLM: Pre-training of Text and Layout for Document Image Understanding” (KDD 2020)
https://arxiv.org/abs/1912.13318
Constructors
| Constructor | Summary |
|---|---|
LayoutLM(NeuralNetworkArchitecture<T> architecture, LayoutLMOptions options, ITokenizer tokenizer, IGradientBasedOptimizer<T, Tensor<T>, Tensor<T>> optimizer, ILossFunction<T> lossFunction) |
Creates a LayoutLM model using native layers for training and inference. |
LayoutLM(NeuralNetworkArchitecture<T> architecture, string onnxModelPath, ITokenizer tokenizer, LayoutLMOptions options, IGradientBasedOptimizer<T, Tensor<T>, Tensor<T>> optimizer, ILossFunction<T> lossFunction) |
Creates a LayoutLM model using a pre-trained ONNX model for inference. |
Properties
| Property | Summary |
|---|---|
DocumentType SupportedDocumentTypes { get; } |
|
IReadOnlyList<LayoutElementType> SupportedElementTypes { get; } |
|
bool RequiresOCR { get; } |
|
int ExpectedImageSize { get; } |
Methods
| Method | Summary |
|---|---|
DocumentLayoutResult<T> DetectLayout(Tensor<T> documentImage) |
|
DocumentLayoutResult<T> DetectLayout(Tensor<T> documentImage, double confidenceThreshold) |
|
ModelMetadata<T> GetModelMetadata() |
|
Tensor<T> EncodeDocument(Tensor<T> documentImage) |
|
string GetModelSummary() |
|
void ValidateInputShape(Tensor<T> documentImage) |