MPLUGDocOwl15<T>

Models & Types · AiDotNet.VisionLanguage.Document

mPLUG-DocOwl 1.5: unified structure learning achieving SOTA on 10 document benchmarks.

Inheritance: NeuralNetworkBase<T> → VisionLanguageModelBase<T> → MPLUGDocOwl15<T>

Implements: IDocumentUnderstandingModel<T>

For Beginners

mPLUG-DocOwl 1.5 is an improved document understanding model with unified structure learning for OCR-free processing. Default values follow the original paper settings.

How It Works

mPLUG-DocOwl 1.5 (Alibaba, 2024) achieves state-of-the-art on 10 document understanding benchmarks through unified structure learning. It learns document layout structure alongside text content in an OCR-free manner, using structure-aware pre-training objectives that teach the model to understand tables, lists, headers, and hierarchical document organization.

References:

  • Paper: “mPLUG-DocOwl 1.5: Unified Structure Learning for OCR-free Document Understanding” (Alibaba, 2024)

Example

// Create an mPLUG-DocOwl 1.5 model for structure-aware document understanding
// with unified structure learning achieving SOTA on 10 benchmarks
var architecture = new NeuralNetworkArchitecture<double>(
    inputType: InputType.TwoDimensional,
    taskType: NeuralNetworkTaskType.ImageClassification,
    inputHeight: 224, inputWidth: 224, inputDepth: 3, outputSize: 512);
            
// ONNX inference mode with pre-trained model
var model = new MPLUGDocOwl15<double>(architecture, "mplugdocowl15.onnx");
            
// Training mode with native layers
var trainModel = new MPLUGDocOwl15<double>(architecture, new MPLUGDocOwl15Options());

Constructors

Constructor Summary
MPLUGDocOwl15(NeuralNetworkArchitecture<T> architecture, MPLUGDocOwl15Options options, IGradientBasedOptimizer<T, Tensor<T>, Tensor<T>> optimizer)
MPLUGDocOwl15(NeuralNetworkArchitecture<T> architecture, string modelPath, MPLUGDocOwl15Options options)

Properties

Property Summary
bool IsOcrFree { get; }
int DecoderEmbeddingDim { get; }
int EmbeddingDimension { get; }
int MaxGenerationLength { get; }

Methods

Method Summary
ModelMetadata<T> GetModelMetadata()
Tensor<T> AnswerDocumentQuestion(Tensor<T> documentImage, string question)
Tensor<T> EncodeImage(Tensor<T> image)
Tensor<T> ExtractText(Tensor<T> documentImage)
Tensor<T> GenerateFromImage(Tensor<T> image, string prompt) Generates text from a document image using mPLUG-DocOwl 1.5’s unified structure learning.