Fix torch device issues (#20584)

Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>

src/transformers/models/conditional_detr/image_processing_conditional_detr.py

@@ -1290,7 +1290,7 @@ class ConditionalDetrImageProcessor(BaseImageProcessor):
 
         # and from relative [0, 1] to absolute [0, height] coordinates
         img_h, img_w = target_sizes.unbind(1)
-        scale_fct = torch.stack([img_w, img_h, img_w, img_h], dim=1).to(boxes.device)
+        scale_fct = torch.stack([img_w, img_h, img_w, img_h], dim=1)
         boxes = boxes * scale_fct[:, None, :]
 
         results = [{"scores": s, "labels": l, "boxes": b} for s, l, b in zip(scores, labels, boxes)]
@@ -1340,7 +1340,7 @@ class ConditionalDetrImageProcessor(BaseImageProcessor):
             img_w = torch.Tensor([i[1] for i in target_sizes])
         else:
             img_h, img_w = target_sizes.unbind(1)
-        scale_fct = torch.stack([img_w, img_h, img_w, img_h], dim=1)
+        scale_fct = torch.stack([img_w, img_h, img_w, img_h], dim=1).to(boxes.device)
         boxes = boxes * scale_fct[:, None, :]
 
         results = []

src/transformers/models/deformable_detr/image_processing_deformable_detr.py

@@ -1288,7 +1288,7 @@ class DeformableDetrImageProcessor(BaseImageProcessor):
 
         # and from relative [0, 1] to absolute [0, height] coordinates
         img_h, img_w = target_sizes.unbind(1)
-        scale_fct = torch.stack([img_w, img_h, img_w, img_h], dim=1).to(boxes.device)
+        scale_fct = torch.stack([img_w, img_h, img_w, img_h], dim=1)
         boxes = boxes * scale_fct[:, None, :]
 
         results = [{"scores": s, "labels": l, "boxes": b} for s, l, b in zip(scores, labels, boxes)]
@@ -1337,7 +1337,7 @@ class DeformableDetrImageProcessor(BaseImageProcessor):
             img_w = torch.Tensor([i[1] for i in target_sizes])
         else:
             img_h, img_w = target_sizes.unbind(1)
-        scale_fct = torch.stack([img_w, img_h, img_w, img_h], dim=1)
+        scale_fct = torch.stack([img_w, img_h, img_w, img_h], dim=1).to(boxes.device)
         boxes = boxes * scale_fct[:, None, :]
 
         results = []