diff --git a/src/transformers/models/deepseek_vl_hybrid/image_processing_deepseek_vl_hybrid.py b/src/transformers/models/deepseek_vl_hybrid/image_processing_deepseek_vl_hybrid.py
index 93dd31e3aafc..241c12923bdb 100644
--- a/src/transformers/models/deepseek_vl_hybrid/image_processing_deepseek_vl_hybrid.py
+++ b/src/transformers/models/deepseek_vl_hybrid/image_processing_deepseek_vl_hybrid.py
@@ -39,12 +39,7 @@
     valid_images,
     validate_preprocess_arguments,
 )
-from ...utils import (
-    TensorType,
-    filter_out_non_signature_kwargs,
-    is_vision_available,
-    logging,
-)
+from ...utils import TensorType, filter_out_non_signature_kwargs, is_vision_available, logging
 
 
 if is_vision_available():
diff --git a/src/transformers/models/deepseek_vl_hybrid/image_processing_deepseek_vl_hybrid_fast.py b/src/transformers/models/deepseek_vl_hybrid/image_processing_deepseek_vl_hybrid_fast.py
index db9c9ad987c1..c04e006e358d 100644
--- a/src/transformers/models/deepseek_vl_hybrid/image_processing_deepseek_vl_hybrid_fast.py
+++ b/src/transformers/models/deepseek_vl_hybrid/image_processing_deepseek_vl_hybrid_fast.py
@@ -21,6 +21,7 @@
 from typing import Optional, Union
 
 import torch
+from torchvision.transforms.v2 import functional as F
 
 from ...image_processing_utils_fast import (
     BaseImageProcessorFast,
@@ -39,13 +40,7 @@
     pil_torch_interpolation_mapping,
 )
 from ...processing_utils import Unpack
-from ...utils import TensorType, auto_docstring, is_torchvision_v2_available
-
-
-if is_torchvision_v2_available():
-    from torchvision.transforms.v2 import functional as F
-else:
-    from torchvision.transforms import functional as F
+from ...utils import TensorType, auto_docstring
 
 
 class DeepseekVLHybridFastImageProcessorKwargs(DefaultFastImageProcessorKwargs):
diff --git a/src/transformers/models/deepseek_vl_hybrid/modeling_deepseek_vl_hybrid.py b/src/transformers/models/deepseek_vl_hybrid/modeling_deepseek_vl_hybrid.py
index cae509e14d64..d9a85654e901 100644
--- a/src/transformers/models/deepseek_vl_hybrid/modeling_deepseek_vl_hybrid.py
+++ b/src/transformers/models/deepseek_vl_hybrid/modeling_deepseek_vl_hybrid.py
@@ -29,11 +29,7 @@
 from ...modeling_outputs import ModelOutput
 from ...modeling_utils import PreTrainedModel
 from ...processing_utils import Unpack
-from ...utils import (
-    TransformersKwargs,
-    auto_docstring,
-    can_return_tuple,
-)
+from ...utils import TransformersKwargs, auto_docstring, can_return_tuple
 from ..auto import AutoModel
 from .configuration_deepseek_vl_hybrid import DeepseekVLHybridConfig
 
diff --git a/src/transformers/models/deepseek_vl_hybrid/modular_deepseek_vl_hybrid.py b/src/transformers/models/deepseek_vl_hybrid/modular_deepseek_vl_hybrid.py
index e9808b02ce34..18b416a57df2 100644
--- a/src/transformers/models/deepseek_vl_hybrid/modular_deepseek_vl_hybrid.py
+++ b/src/transformers/models/deepseek_vl_hybrid/modular_deepseek_vl_hybrid.py
@@ -16,6 +16,7 @@
 
 import torch
 import torch.nn as nn
+from torchvision.transforms.v2 import functional as F
 
 from ...cache_utils import Cache
 from ...image_processing_utils_fast import (
@@ -53,7 +54,6 @@
     auto_docstring,
     can_return_tuple,
     filter_out_non_signature_kwargs,
-    is_torchvision_v2_available,
     logging,
 )
 from ..auto import CONFIG_MAPPING, AutoConfig, AutoModel
@@ -70,12 +70,6 @@
 from ..sam.modeling_sam import SamLayerNorm, SamVisionNeck
 
 
-if is_torchvision_v2_available():
-    from torchvision.transforms.v2 import functional as F
-else:
-    from torchvision.transforms import functional as F
-
-
 logger = logging.get_logger(__name__)
 
 
diff --git a/src/transformers/models/dpt/image_processing_dpt_fast.py b/src/transformers/models/dpt/image_processing_dpt_fast.py
index 3e80ad7943db..892ddd7c3d6f 100644
--- a/src/transformers/models/dpt/image_processing_dpt_fast.py
+++ b/src/transformers/models/dpt/image_processing_dpt_fast.py
@@ -25,6 +25,7 @@
 from typing import TYPE_CHECKING, Optional, Union
 
 import torch
+from torchvision.transforms.v2 import functional as F
 
 from ...image_processing_base import BatchFeature
 from ...image_processing_utils_fast import BaseImageProcessorFast, DefaultFastImageProcessorKwargs
@@ -39,17 +40,12 @@
     is_torch_tensor,
 )
 from ...processing_utils import Unpack
-from ...utils import TensorType, auto_docstring, is_torchvision_v2_available, requires_backends
+from ...utils import TensorType, auto_docstring, requires_backends
 
 
 if TYPE_CHECKING:
     from ...modeling_outputs import DepthEstimatorOutput
 
-if is_torchvision_v2_available():
-    from torchvision.transforms.v2 import functional as F
-else:
-    from torchvision.transforms import functional as F
-
 
 class DPTFastImageProcessorKwargs(DefaultFastImageProcessorKwargs):
     """
diff --git a/src/transformers/models/dpt/modular_dpt.py b/src/transformers/models/dpt/modular_dpt.py
index 32ca94a2d43f..34eb08f39b68 100644
--- a/src/transformers/models/dpt/modular_dpt.py
+++ b/src/transformers/models/dpt/modular_dpt.py
@@ -32,7 +32,6 @@
 from ...utils import (
     TensorType,
     auto_docstring,
-    is_torchvision_v2_available,
     requires_backends,
 )
 from ..beit.image_processing_beit_fast import BeitImageProcessorFast
@@ -41,10 +40,7 @@
 if TYPE_CHECKING:
     from ...modeling_outputs import DepthEstimatorOutput
 
-if is_torchvision_v2_available():
-    from torchvision.transforms.v2 import functional as F
-else:
-    from torchvision.transforms import functional as F
+from torchvision.transforms.v2 import functional as F
 
 
 def get_resize_output_image_size(
diff --git a/src/transformers/models/llava_onevision/image_processing_llava_onevision_fast.py b/src/transformers/models/llava_onevision/image_processing_llava_onevision_fast.py
index 4392d64e9ebf..11872cb67bf3 100644
--- a/src/transformers/models/llava_onevision/image_processing_llava_onevision_fast.py
+++ b/src/transformers/models/llava_onevision/image_processing_llava_onevision_fast.py
@@ -22,6 +22,7 @@
 from typing import Optional, Union
 
 import torch
+from torchvision.transforms.v2 import functional as F
 
 from ...image_processing_utils import BatchFeature, get_patch_output_size, select_best_resolution
 from ...image_processing_utils_fast import (
@@ -41,13 +42,7 @@
     get_image_size,
 )
 from ...processing_utils import Unpack
-from ...utils import TensorType, auto_docstring, is_torchvision_v2_available
-
-
-if is_torchvision_v2_available():
-    from torchvision.transforms.v2 import functional as F
-else:
-    from torchvision.transforms import functional as F
+from ...utils import TensorType, auto_docstring
 
 
 class LlavaOnevisionFastImageProcessorKwargs(DefaultFastImageProcessorKwargs):
diff --git a/src/transformers/models/llava_onevision/modular_llava_onevision.py b/src/transformers/models/llava_onevision/modular_llava_onevision.py
index ec2304e09dd1..b4f64dee8e04 100644
--- a/src/transformers/models/llava_onevision/modular_llava_onevision.py
+++ b/src/transformers/models/llava_onevision/modular_llava_onevision.py
@@ -18,6 +18,7 @@
 
 import torch
 from torch import nn
+from torchvision.transforms.v2 import functional as F
 
 from transformers.models.llava_next.image_processing_llava_next_fast import LlavaNextImageProcessorFast
 from transformers.models.llava_next_video.modeling_llava_next_video import (
@@ -50,16 +51,10 @@
     TensorType,
     auto_docstring,
     can_return_tuple,
-    is_torchvision_v2_available,
     logging,
 )
 
 
-if is_torchvision_v2_available():
-    from torchvision.transforms.v2 import functional as F
-else:
-    from torchvision.transforms import functional as F
-
 logger = logging.get_logger(__name__)
 
 
diff --git a/src/transformers/models/owlv2/image_processing_owlv2_fast.py b/src/transformers/models/owlv2/image_processing_owlv2_fast.py
index c17a45b6e427..417fc800ea88 100644
--- a/src/transformers/models/owlv2/image_processing_owlv2_fast.py
+++ b/src/transformers/models/owlv2/image_processing_owlv2_fast.py
@@ -23,6 +23,7 @@
 from typing import TYPE_CHECKING, Optional, Union
 
 import torch
+from torchvision.transforms.v2 import functional as F
 
 from ...image_processing_utils_fast import BaseImageProcessorFast, BatchFeature, DefaultFastImageProcessorKwargs
 from ...image_transforms import center_to_corners_format, group_images_by_shape, reorder_images
@@ -35,16 +36,10 @@
     SizeDict,
 )
 from ...processing_utils import Unpack
-from ...utils import TensorType, auto_docstring, is_torchvision_v2_available
+from ...utils import TensorType, auto_docstring
 from .image_processing_owlv2 import _scale_boxes, box_iou
 
 
-if is_torchvision_v2_available():
-    from torchvision.transforms.v2 import functional as F
-else:
-    from torchvision.transforms import functional as F
-
-
 if TYPE_CHECKING:
     from .modeling_owlv2 import Owlv2ObjectDetectionOutput
 
diff --git a/src/transformers/models/owlv2/modular_owlv2.py b/src/transformers/models/owlv2/modular_owlv2.py
index 2e6d917a791a..66acd2088399 100644
--- a/src/transformers/models/owlv2/modular_owlv2.py
+++ b/src/transformers/models/owlv2/modular_owlv2.py
@@ -18,6 +18,7 @@
 from typing import Optional, Union
 
 import torch
+from torchvision.transforms.v2 import functional as F
 
 from ...image_processing_utils_fast import (
     BaseImageProcessorFast,
@@ -37,17 +38,10 @@
 from ...utils import (
     TensorType,
     auto_docstring,
-    is_torchvision_v2_available,
 )
 from ..owlvit.image_processing_owlvit_fast import OwlViTImageProcessorFast
 
 
-if is_torchvision_v2_available():
-    from torchvision.transforms.v2 import functional as F
-else:
-    from torchvision.transforms import functional as F
-
-
 class Owlv2FastImageProcessorKwargs(DefaultFastImageProcessorKwargs): ...
 
 
diff --git a/src/transformers/models/rt_detr/image_processing_rt_detr_fast.py b/src/transformers/models/rt_detr/image_processing_rt_detr_fast.py
index 68c5497b0205..1dfffd31082c 100644
--- a/src/transformers/models/rt_detr/image_processing_rt_detr_fast.py
+++ b/src/transformers/models/rt_detr/image_processing_rt_detr_fast.py
@@ -8,6 +8,7 @@
 from typing import Any, Optional, Union
 
 import torch
+from torchvision.transforms.v2 import functional as F
 
 from ...image_processing_utils import BatchFeature
 from ...image_processing_utils_fast import (
@@ -36,12 +37,6 @@
 from .image_processing_rt_detr import get_size_with_aspect_ratio
 
 
-if is_torchvision_v2_available():
-    from torchvision.transforms.v2 import functional as F
-else:
-    from torchvision.transforms import functional as F
-
-
 class RTDetrFastImageProcessorKwargs(DefaultFastImageProcessorKwargs):
     r"""
     format (`str`, *optional*, defaults to `AnnotationFormat.COCO_DETECTION`):
diff --git a/src/transformers/models/rt_detr/modular_rt_detr.py b/src/transformers/models/rt_detr/modular_rt_detr.py
index 760e4a6675cf..61bd055144f0 100644
--- a/src/transformers/models/rt_detr/modular_rt_detr.py
+++ b/src/transformers/models/rt_detr/modular_rt_detr.py
@@ -2,6 +2,7 @@
 from typing import Optional, Union
 
 import torch
+from torchvision.transforms.v2 import functional as F
 
 from transformers.models.detr.image_processing_detr_fast import DetrFastImageProcessorKwargs, DetrImageProcessorFast
 
@@ -22,18 +23,11 @@
 from ...processing_utils import Unpack
 from ...utils import (
     TensorType,
-    is_torchvision_v2_available,
     logging,
     requires_backends,
 )
 
 
-if is_torchvision_v2_available():
-    from torchvision.transforms.v2 import functional as F
-else:
-    from torchvision.transforms import functional as F
-
-
 logger = logging.get_logger(__name__)
 
 SUPPORTED_ANNOTATION_FORMATS = (AnnotationFormat.COCO_DETECTION,)
diff --git a/src/transformers/models/sam2_video/modeling_sam2_video.py b/src/transformers/models/sam2_video/modeling_sam2_video.py
index caa07d1f63b5..79d5b015f889 100644
--- a/src/transformers/models/sam2_video/modeling_sam2_video.py
+++ b/src/transformers/models/sam2_video/modeling_sam2_video.py
@@ -39,10 +39,7 @@
 from ...modeling_utils import ALL_ATTENTION_FUNCTIONS, PreTrainedModel
 from ...processing_utils import Unpack
 from ...pytorch_utils import compile_compatible_method_lru_cache
-from ...utils import (
-    ModelOutput,
-    auto_docstring,
-)
+from ...utils import ModelOutput, auto_docstring
 from ...utils.generic import OutputRecorder, TransformersKwargs
 from ..auto import AutoModel
 from .configuration_sam2_video import Sam2VideoConfig, Sam2VideoMaskDecoderConfig, Sam2VideoPromptEncoderConfig
diff --git a/src/transformers/models/sam2_video/modular_sam2_video.py b/src/transformers/models/sam2_video/modular_sam2_video.py
index fa0d6c21d5e6..b95a9f778251 100644
--- a/src/transformers/models/sam2_video/modular_sam2_video.py
+++ b/src/transformers/models/sam2_video/modular_sam2_video.py
@@ -36,8 +36,6 @@
 from ...utils import (
     ModelOutput,
     auto_docstring,
-    is_torchvision_available,
-    is_torchvision_v2_available,
     logging,
 )
 from ...utils.generic import OutputRecorder, TransformersKwargs
@@ -59,12 +57,6 @@
 from ..sam2.processing_sam2 import Sam2Processor
 
 
-if is_torchvision_available() and is_torchvision_v2_available():
-    from torchvision.transforms.v2 import functional as F
-else:
-    from torchvision.transforms import functional as F
-
-
 logger = logging.get_logger(__name__)
 
 
diff --git a/src/transformers/models/segformer/image_processing_segformer_fast.py b/src/transformers/models/segformer/image_processing_segformer_fast.py
index dc18283136e1..11dfa3c42ab1 100644
--- a/src/transformers/models/segformer/image_processing_segformer_fast.py
+++ b/src/transformers/models/segformer/image_processing_segformer_fast.py
@@ -22,6 +22,7 @@
 from typing import Optional, Union
 
 import torch
+from torchvision.transforms.v2 import functional as F
 
 from ...image_processing_utils import BatchFeature
 from ...image_processing_utils_fast import (
@@ -40,13 +41,7 @@
     is_torch_tensor,
 )
 from ...processing_utils import Unpack
-from ...utils import TensorType, auto_docstring, is_torchvision_v2_available
-
-
-if is_torchvision_v2_available():
-    from torchvision.transforms.v2 import functional as F
-else:
-    from torchvision.transforms import functional as F
+from ...utils import TensorType, auto_docstring
 
 
 class SegformerFastImageProcessorKwargs(DefaultFastImageProcessorKwargs):
@@ -135,9 +130,7 @@ def _preprocess_image_like_inputs(
                     "do_normalize": False,
                     "do_rescale": False,
                     # Nearest interpolation is used for segmentation maps instead of BILINEAR.
-                    "interpolation": F.InterpolationMode.NEAREST_EXACT
-                    if is_torchvision_v2_available()
-                    else F.InterpolationMode.NEAREST,
+                    "interpolation": F.InterpolationMode.NEAREST_EXACT,
                 }
             )
             processed_segmentation_maps = self._preprocess(
diff --git a/src/transformers/models/segformer/modular_segformer.py b/src/transformers/models/segformer/modular_segformer.py
index 341e6949d8b7..831d046fd9a7 100644
--- a/src/transformers/models/segformer/modular_segformer.py
+++ b/src/transformers/models/segformer/modular_segformer.py
@@ -17,6 +17,7 @@
 from typing import Optional, Union
 
 import torch
+from torchvision.transforms.v2 import functional as F
 
 from transformers.models.beit.image_processing_beit_fast import BeitFastImageProcessorKwargs, BeitImageProcessorFast
 
@@ -36,16 +37,9 @@
 from ...processing_utils import Unpack
 from ...utils import (
     TensorType,
-    is_torchvision_v2_available,
 )
 
 
-if is_torchvision_v2_available():
-    from torchvision.transforms.v2 import functional as F
-else:
-    from torchvision.transforms import functional as F
-
-
 class SegformerFastImageProcessorKwargs(BeitFastImageProcessorKwargs):
     pass
 
@@ -96,9 +90,7 @@ def _preprocess_image_like_inputs(
                     "do_normalize": False,
                     "do_rescale": False,
                     # Nearest interpolation is used for segmentation maps instead of BILINEAR.
-                    "interpolation": F.InterpolationMode.NEAREST_EXACT
-                    if is_torchvision_v2_available()
-                    else F.InterpolationMode.NEAREST,
+                    "interpolation": F.InterpolationMode.NEAREST_EXACT,
                 }
             )
             processed_segmentation_maps = self._preprocess(
diff --git a/src/transformers/utils/import_utils.py b/src/transformers/utils/import_utils.py
index 430153690c79..07519ee865ac 100644
--- a/src/transformers/utils/import_utils.py
+++ b/src/transformers/utils/import_utils.py
@@ -400,11 +400,7 @@ def is_torchvision_available() -> bool:
 
 
 def is_torchvision_v2_available() -> bool:
-    if not is_torchvision_available():
-        return False
-
-    # NOTE: We require torchvision>=0.15 as v2 transforms are available from this version: https://pytorch.org/vision/stable/transforms.html#v1-or-v2-which-one-should-i-use
-    return version.parse(_torchvision_version) >= version.parse("0.15")
+    return is_torchvision_available()
 
 
 def is_galore_torch_available() -> Union[tuple[bool, str], bool]: