fix type hints and spelling mistake in generalized_rcnn and poolers (#2550)

* fix type hints and move degenerate boxes to a function in torchvision.models.detection.generalized_rcnn

* format code

* format code

* changed to static method

* revert imports

* changed to method

* revert procedure for degenerating boxes

torchvision/models/detection/generalized_rcnn.py

@@ -4,6 +4,7 @@ Implements the Generalized R-CNN framework
 """
 
 from collections import OrderedDict
+from typing import Union
 import torch
 from torch import nn
 import warnings
@@ -35,7 +36,7 @@ class GeneralizedRCNN(nn.Module):
 
     @torch.jit.unused
     def eager_outputs(self, losses, detections):
-        # type: (Dict[str, Tensor], List[Dict[str, Tensor]]) -> Tuple[Dict[str, Tensor], List[Dict[str, Tensor]]]
+        # type: (Dict[str, Tensor], List[Dict[str, Tensor]]) -> Union[Dict[str, Tensor], List[Dict[str, Tensor]]]
         if self.training:
             return losses
 
@@ -85,11 +86,11 @@ class GeneralizedRCNN(nn.Module):
                 boxes = target["boxes"]
                 degenerate_boxes = boxes[:, 2:] <= boxes[:, :2]
                 if degenerate_boxes.any():
-                    # print the first degenrate box
+                    # print the first degenerate box
                     bb_idx = degenerate_boxes.any(dim=1).nonzero().view(-1)[0]
                     degen_bb: List[float] = boxes[bb_idx].tolist()
                     raise ValueError("All bounding boxes should have positive height and width."
-                                     " Found invaid box {} for target at index {}."
+                                     " Found invalid box {} for target at index {}."
                                      .format(degen_bb, target_idx))
 
         features = self.backbone(images.tensors)

torchvision/ops/poolers.py

 # Copyright (c) Facebook, Inc. and its affiliates. All Rights Reserved.
+from typing import Union
+
 import torch
 import torch.nn.functional as F
 from torch import nn, Tensor
@@ -119,7 +121,7 @@ class MultiScaleRoIAlign(nn.Module):
     def __init__(
         self,
         featmap_names: List[str],
-        output_size: List[int],
+        output_size: Union[int, Tuple[int], List[int]],
         sampling_ratio: int,
     ):
         super(MultiScaleRoIAlign, self).__init__()