fix: address deprecation warnings of dependencies (#2237)

* switch to dtype instead of torch_dtype

Signed-off-by: Michele Dolfi <dol@zurich.ibm.com>

* set __check_model__ to avoid deprecation warnings

Signed-off-by: Michele Dolfi <dol@zurich.ibm.com>

* remove dataloaders warnings in easyocr

Signed-off-by: Michele Dolfi <dol@zurich.ibm.com>

* suppress with option

Signed-off-by: Michele Dolfi <dol@zurich.ibm.com>

---------

Signed-off-by: Michele Dolfi <dol@zurich.ibm.com>

docling/datamodel/pipeline_options.py

@@ -135,6 +135,8 @@ class EasyOcrOptions(OcrOptions):
     recog_network: Optional[str] = "standard"
     download_enabled: bool = True
 
+    suppress_mps_warnings: bool = True
+
     model_config = ConfigDict(
         extra="forbid",
         protected_namespaces=(),

docling/models/easyocr_model.py

@@ -78,14 +78,17 @@ class EasyOcrModel(BaseOcrModel):
                 download_enabled = False
                 model_storage_directory = str(artifacts_path / self._model_repo_folder)
 
-            self.reader = easyocr.Reader(
+            with warnings.catch_warnings():
-                lang_list=self.options.lang,
+                if self.options.suppress_mps_warnings:
-                gpu=use_gpu,
+                    warnings.filterwarnings("ignore", message=".*pin_memory.*MPS.*")
-                model_storage_directory=model_storage_directory,
+                self.reader = easyocr.Reader(
-                recog_network=self.options.recog_network,
+                    lang_list=self.options.lang,
-                download_enabled=download_enabled,
+                    gpu=use_gpu,
-                verbose=False,
+                    model_storage_directory=model_storage_directory,
-            )
+                    recog_network=self.options.recog_network,
+                    download_enabled=download_enabled,
+                    verbose=False,
+                )
 
     @staticmethod
     def download_models(
@@ -147,7 +150,14 @@ class EasyOcrModel(BaseOcrModel):
                             scale=self.scale, cropbox=ocr_rect
                         )
                         im = numpy.array(high_res_image)
-                        result = self.reader.readtext(im)
+
+                        with warnings.catch_warnings():
+                            if self.options.suppress_mps_warnings:
+                                warnings.filterwarnings(
+                                    "ignore", message=".*pin_memory.*MPS.*"
+                                )
+
+                            result = self.reader.readtext(im)
 
                         del high_res_image
                         del im

docling/models/picture_description_vlm_model.py

@@ -67,7 +67,7 @@ class PictureDescriptionVlmModel(
                 self.model = AutoModelForImageTextToText.from_pretrained(
                     artifacts_path,
                     device_map=self.device,
-                    torch_dtype=torch.bfloat16,
+                    dtype=torch.bfloat16,
                     _attn_implementation=(
                         "flash_attention_2"
                         if self.device.startswith("cuda")

docling/models/vlm_models_inline/hf_transformers_model.py

@@ -112,7 +112,7 @@ class HuggingFaceTransformersVlmModel(BaseVlmPageModel, HuggingFaceModelDownload
             self.vlm_model = model_cls.from_pretrained(
                 artifacts_path,
                 device_map=self.device,
-                torch_dtype=self.vlm_options.torch_dtype,
+                dtype=self.vlm_options.torch_dtype,
                 _attn_implementation=(
                     "flash_attention_2"
                     if self.device.startswith("cuda")

docling/models/vlm_models_inline/nuextract_transformers_model.py

@@ -144,7 +144,7 @@ class NuExtractTransformersModel(BaseVlmModel, HuggingFaceModelDownloadMixin):
             self.vlm_model = AutoModelForImageTextToText.from_pretrained(
                 artifacts_path,
                 device_map=self.device,
-                torch_dtype=self.vlm_options.torch_dtype,
+                dtype=self.vlm_options.torch_dtype,
                 _attn_implementation=(
                     "flash_attention_2"
                     if self.device.startswith("cuda")

docling/pipeline/extraction_vlm_pipeline.py

@@ -194,6 +194,9 @@ class ExtractionVlmPipeline(BaseExtractionPipeline):
             class ExtractionTemplateFactory(ModelFactory[template]):  # type: ignore
                 __use_examples__ = True  # prefer Field(examples=...) when present
                 __use_defaults__ = True  # use field defaults instead of random values
+                __check_model__ = (
+                    True  # setting the value to avoid deprecation warnings
+                )
 
             return ExtractionTemplateFactory.build().model_dump_json(indent=2)  # type: ignore
         else: