Model Catalog

from docling.datamodel.pipeline_options import LayoutOptions
from docling.datamodel.layout_model_specs import DOCLING_LAYOUT_HERON

# Use Heron layout model (default)
layout_options = LayoutOptions(
    model_spec=DOCLING_LAYOUT_HERON
)

from docling.datamodel.pipeline_options import (
    TesseractOcrOptions,
    RapidOcrOptions
)

# Tesseract with multiple languages
ocr_options = TesseractOcrOptions(
    lang=["eng", "deu"]  # English and German
)

# RapidOCR with GPU acceleration
ocr_options = RapidOcrOptions(
    backend="torch",  # GPU-accelerated
    lang=["en"]
)

from docling.datamodel.pipeline_options import (
    TableStructureOptions,
    TableFormerMode
)

# Use accurate mode for best quality
table_options = TableStructureOptions(
    mode=TableFormerMode.ACCURATE,
    do_cell_matching=True  # Align cells with content
)

from docling.models.stages.picture_classifier.document_picture_classifier import (
    DocumentPictureClassifierOptions
)

# Use default picture classifier
classifier_options = DocumentPictureClassifierOptions.from_preset(
    "document_figure_classifier_v2"
)

from docling.datamodel.pipeline_options import VlmConvertOptions

# Use SmolDocling with auto-selected engine
options = VlmConvertOptions.from_preset("smoldocling")

# Force specific engine
from docling.datamodel.vlm_engine_options import MlxVlmEngineOptions

options = VlmConvertOptions.from_preset(
    "smoldocling",
    engine_options=MlxVlmEngineOptions()
)

<document>
  <section_header>Introduction</section_header>
  <text>This is a paragraph...</text>
  <table>
    <row><cell>Data</cell></row>
  </table>
</document>

# Introduction

This is a paragraph...

| Column 1 | Column 2 |
|----------|----------|
| Data     | Data     |

from docling.datamodel.pipeline_options import PictureDescriptionVlmOptions

# Use Granite Vision for detailed descriptions
options = PictureDescriptionVlmOptions.from_preset("granite_vision")

from docling.datamodel.pipeline_options import CodeFormulaVlmOptions

# Use specialized CodeFormulaV2 model
options = CodeFormulaVlmOptions.from_preset("codeformulav2")