From b9c16982635bad806e1ca4ef0cc4843c71ed2ffa Mon Sep 17 00:00:00 2001 From: Michele Dolfi Date: Mon, 2 Jun 2025 10:40:06 +0200 Subject: [PATCH] rename to specs Signed-off-by: Michele Dolfi --- docling/cli/main.py | 2 +- docling/datamodel/pipeline_options.py | 2 +- .../{vlm_model_spec.py => vlm_model_specs.py} | 10 +++++++--- docling/utils/model_downloader.py | 2 +- docs/examples/compare_vlm_models.py | 2 +- docs/examples/minimal_vlm_pipeline.py | 2 +- 6 files changed, 12 insertions(+), 8 deletions(-) rename docling/datamodel/{vlm_model_spec.py => vlm_model_specs.py} (95%) diff --git a/docling/cli/main.py b/docling/cli/main.py index 16810778..083f53b2 100644 --- a/docling/cli/main.py +++ b/docling/cli/main.py @@ -47,7 +47,7 @@ from docling.datamodel.pipeline_options import ( VlmPipelineOptions, ) from docling.datamodel.settings import settings -from docling.datamodel.vlm_model_spec import ( +from docling.datamodel.vlm_model_specs import ( GRANITE_VISION_OLLAMA, GRANITE_VISION_TRANSFORMERS, SMOLDOCLING_MLX, diff --git a/docling/datamodel/pipeline_options.py b/docling/datamodel/pipeline_options.py index ba8a39f9..cde5cd33 100644 --- a/docling/datamodel/pipeline_options.py +++ b/docling/datamodel/pipeline_options.py @@ -19,7 +19,7 @@ from docling.datamodel.pipeline_options_vlm_model import ( InlineVlmOptions, ResponseFormat, ) -from docling.datamodel.vlm_model_spec import ( +from docling.datamodel.vlm_model_specs import ( GRANITE_VISION_OLLAMA as granite_vision_vlm_ollama_conversion_options, GRANITE_VISION_TRANSFORMERS as granite_vision_vlm_conversion_options, SMOLDOCLING_MLX as smoldocling_vlm_mlx_conversion_options, diff --git a/docling/datamodel/vlm_model_spec.py b/docling/datamodel/vlm_model_specs.py similarity index 95% rename from docling/datamodel/vlm_model_spec.py rename to docling/datamodel/vlm_model_specs.py index 5a921f90..663cbd73 100644 --- a/docling/datamodel/vlm_model_spec.py +++ b/docling/datamodel/vlm_model_specs.py @@ -44,10 +44,14 @@ SMOLDOCLING_TRANSFORMERS = InlineVlmOptions( # GraniteVision GRANITE_VISION_TRANSFORMERS = InlineVlmOptions( repo_id="ibm-granite/granite-vision-3.2-2b", - prompt="Convert this page to markdown. Do not miss any text and only output the bare MarkDown!", + prompt="Convert this page to markdown. Do not miss any text and only output the bare markdown!", response_format=ResponseFormat.MARKDOWN, inference_framework=InferenceFramework.TRANSFORMERS_VISION2SEQ, - supported_devices=[AcceleratorDevice.CPU, AcceleratorDevice.CUDA], + supported_devices=[ + AcceleratorDevice.CPU, + AcceleratorDevice.CUDA, + AcceleratorDevice.MPS, + ], scale=2.0, temperature=0.0, ) @@ -55,7 +59,7 @@ GRANITE_VISION_TRANSFORMERS = InlineVlmOptions( GRANITE_VISION_OLLAMA = ApiVlmOptions( url=AnyUrl("http://localhost:11434/v1/chat/completions"), params={"model": "granite3.2-vision:2b"}, - prompt="Convert this page to markdown. Do not miss any text and only output the bare MarkDown!", + prompt="Convert this page to markdown. Do not miss any text and only output the bare markdown!", scale=1.0, timeout=120, response_format=ResponseFormat.MARKDOWN, diff --git a/docling/utils/model_downloader.py b/docling/utils/model_downloader.py index 8a8e1eb4..f8237fbc 100644 --- a/docling/utils/model_downloader.py +++ b/docling/utils/model_downloader.py @@ -7,7 +7,7 @@ from docling.datamodel.pipeline_options import ( smolvlm_picture_description, ) from docling.datamodel.settings import settings -from docling.datamodel.vlm_model_spec import ( +from docling.datamodel.vlm_model_specs import ( SMOLDOCLING_MLX, SMOLDOCLING_TRANSFORMERS, ) diff --git a/docs/examples/compare_vlm_models.py b/docs/examples/compare_vlm_models.py index 0ab237ab..5584a699 100644 --- a/docs/examples/compare_vlm_models.py +++ b/docs/examples/compare_vlm_models.py @@ -16,7 +16,7 @@ from docling.datamodel.base_models import InputFormat from docling.datamodel.pipeline_options import ( VlmPipelineOptions, ) -from docling.datamodel.vlm_model_spec import ( +from docling.datamodel.vlm_model_specs import ( GEMMA3_12B_MLX, GRANITE_VISION_OLLAMA, GRANITE_VISION_TRANSFORMERS, diff --git a/docs/examples/minimal_vlm_pipeline.py b/docs/examples/minimal_vlm_pipeline.py index b0c27eb4..0ace6af9 100644 --- a/docs/examples/minimal_vlm_pipeline.py +++ b/docs/examples/minimal_vlm_pipeline.py @@ -2,7 +2,7 @@ from docling.datamodel.base_models import InputFormat from docling.datamodel.pipeline_options import ( VlmPipelineOptions, ) -from docling.datamodel.vlm_model_spec import SMOLDOCLING_MLX +from docling.datamodel.vlm_model_specs import SMOLDOCLING_MLX from docling.document_converter import DocumentConverter, PdfFormatOption from docling.pipeline.vlm_pipeline import VlmPipeline