VLM pipeline with SmolDocling

In [ ]:

from docling.datamodel import vlm_model_specs
from docling.datamodel.base_models import InputFormat
from docling.datamodel.pipeline_options import (
    VlmPipelineOptions,
)
from docling.document_converter import DocumentConverter, PdfFormatOption
from docling.pipeline.vlm_pipeline import VlmPipeline

from docling.datamodel import vlm_model_specs from docling.datamodel.base_models import InputFormat from docling.datamodel.pipeline_options import ( VlmPipelineOptions, ) from docling.document_converter import DocumentConverter, PdfFormatOption from docling.pipeline.vlm_pipeline import VlmPipeline

In [ ]:

source = "https://arxiv.org/pdf/2501.17887"

source = "https://arxiv.org/pdf/2501.17887"

USING SIMPLE DEFAULT VALUES¶

• SmolDocling model
• Using the transformers framework

In [ ]:

converter = DocumentConverter(
    format_options={
        InputFormat.PDF: PdfFormatOption(
            pipeline_cls=VlmPipeline,
        ),
    }
)

converter = DocumentConverter( format_options={ InputFormat.PDF: PdfFormatOption( pipeline_cls=VlmPipeline, ), } )

In [ ]:

doc = converter.convert(source=source).document

doc = converter.convert(source=source).document

In [ ]:

print(doc.export_to_markdown())

print(doc.export_to_markdown())

USING MACOS MPS ACCELERATOR¶

For more options see the compare_vlm_models.py example.

In [ ]:

pipeline_options = VlmPipelineOptions(
    vlm_options=vlm_model_specs.SMOLDOCLING_MLX,
)

pipeline_options = VlmPipelineOptions( vlm_options=vlm_model_specs.SMOLDOCLING_MLX, )

In [ ]:

converter = DocumentConverter(
    format_options={
        InputFormat.PDF: PdfFormatOption(
            pipeline_cls=VlmPipeline,
            pipeline_options=pipeline_options,
        ),
    }
)

converter = DocumentConverter( format_options={ InputFormat.PDF: PdfFormatOption( pipeline_cls=VlmPipeline, pipeline_options=pipeline_options, ), } )

In [ ]:

doc = converter.convert(source=source).document

doc = converter.convert(source=source).document

In [ ]:

print(doc.export_to_markdown())

print(doc.export_to_markdown())