feat: transformers to v5

docling_ibm_models/code_formula_model/code_formula_predictor.py

@@ -254,7 +254,7 @@ def predict(
         )
 
         if self._device == "cpu":
-            output_ids_list = self._model.generate(
+            output_ids_list = self._model.generate(  # type: ignore[misc]
                 input_ids=prompt_ids,
                 attention_mask=attention_mask,
                 images=images_tensor,
@@ -267,7 +267,7 @@ def predict(
             )
         else:
             with torch.autocast(device_type=self._device, dtype=torch.bfloat16):
-                output_ids_list = self._model.generate(
+                output_ids_list = self._model.generate(  # type: ignore[misc]
                     prompt_ids,
                     images=images_tensor,
                     do_sample=do_sample,

docling_ibm_models/code_formula_model/models/sam_opt_image_processor.py

@@ -1,10 +1,9 @@
 from PIL import Image
 from torchvision.transforms import functional as F
-from transformers import AutoImageProcessor
-from transformers.image_processing_utils import ImageProcessingMixin
+from transformers import AutoImageProcessor, BaseImageProcessor
 
 
-class SamOptImageProcessor(ImageProcessingMixin):
+class SamOptImageProcessor(BaseImageProcessor):
 
     def __init__(self, size=(1024, 1024), mean=None, std=None, **kwargs):
         super().__init__(**kwargs)

docling_ibm_models/tableformer_v2/model.py

@@ -18,7 +18,13 @@
 import torch
 import torch.nn as nn
 from torchvision.models import efficientnet_v2_s
-from transformers import AutoConfig, AutoModel, PretrainedConfig, PreTrainedModel
+from transformers import (
+    AutoConfig,
+    AutoModel,
+    GenerationMixin,
+    PretrainedConfig,
+    PreTrainedModel,
+)
 from transformers.modeling_outputs import ModelOutput
 
 _log = logging.getLogger(__name__)
@@ -539,7 +545,7 @@ def __init__(
 # =============================================================================
 
 
-class TableFormerV2(PreTrainedModel):
+class TableFormerV2(PreTrainedModel, GenerationMixin):
     r"""
     TableFormerV2: CPU-optimized model for table structure recognition (inference only).
 
@@ -591,6 +597,8 @@ def __init__(self, config: TableFormerV2Config):
 
         self.data_cells = config.data_cells
 
+        self.post_init()
+
     def _is_profiling_enabled(self) -> bool:
         r"""
         Check if profiling is enabled by checking if AggProfiler has cycles.
@@ -779,7 +787,7 @@ def forward(
             past_key_values=present_kv,
         )
 
-    def prepare_inputs_for_generation(
+    def prepare_inputs_for_generation(  # type: ignore[override]
         self,
         input_ids: torch.Tensor,
         past: Optional[Tuple] = None,
@@ -795,7 +803,7 @@ def prepare_inputs_for_generation(
             "use_cache": True,
         }
 
-    def generate(
+    def generate(  # type: ignore[override]
         self,
         images: torch.Tensor,
         tokenizer,

pyproject.toml

@@ -39,7 +39,7 @@ dependencies = [
   'safetensors[torch] (>=0.4.3,<1)',
   'pydantic (>=2.0.0,<3.0.0)',
   'docling-core (>=2.19.0,<3.0.0)',
-  'transformers (>=4.42.0,<5.0.0)',
+  'transformers (>=4.42.0,<6.0.0,!=5.0.*,!=5.1.*,!=5.2.*,!=5.3.*)',
   'numpy (>=1.24.4,<3.0.0)',
   "rtree>=1.0.0",
   'accelerate (>=1.2.1,<2.0.0)',