fix Load DETR-family panoptic masks with load_image_as_tensor (#50234)

Load DETR-family panoptic masks with load_image_as_tensor

Author: dhekimian

src/transformers/models/conditional_detr/image_processing_conditional_detr.py

@@ -19,7 +19,6 @@
 import numpy as np
 import torch
 from torch import nn
-from torchvision.io import read_image
 from torchvision.transforms.v2 import functional as tvF
 
 from ...image_processing_backends import TorchvisionBackend
@@ -45,6 +44,7 @@
     get_image_size,
     get_image_size_for_max_height_width,
     get_max_height_width,
+    load_image_as_tensor,
     validate_annotations,
 )
 from ...processing_utils import ImagesKwargs, Unpack
@@ -394,7 +394,7 @@ def prepare_coco_panoptic_annotation(
     new_target["orig_size"] = torch.as_tensor([image_height, image_width], dtype=torch.int64, device=image.device)
 
     if "segments_info" in target:
-        masks = read_image(annotation_path).permute(1, 2, 0).to(dtype=torch.int32, device=image.device)
+        masks = load_image_as_tensor(str(annotation_path)).permute(1, 2, 0).to(dtype=torch.int32, device=image.device)
         masks = rgb_to_id(masks)
 
         ids = torch.as_tensor([segment_info["id"] for segment_info in target["segments_info"]], device=image.device)