Add non-TS'able _resize_image_and_masks variant with less tensor ops (#7592)

Signed-off-by: Edward Z. Yang <ezyang@meta.com>

torchvision/models/detection/transform.py

@@ -24,8 +24,8 @@ def _fake_cast_onnx(v: Tensor) -> float:
 
 def _resize_image_and_masks(
     image: Tensor,
-    self_min_size: float,
-    self_max_size: float,
+    self_min_size: int,
+    self_max_size: int,
     target: Optional[Dict[str, Tensor]] = None,
     fixed_size: Optional[Tuple[int, int]] = None,
 ) -> Tuple[Tensor, Optional[Dict[str, Tensor]]]:
@@ -40,14 +40,24 @@ def _resize_image_and_masks(
     if fixed_size is not None:
         size = [fixed_size[1], fixed_size[0]]
     else:
-        min_size = torch.min(im_shape).to(dtype=torch.float32)
-        max_size = torch.max(im_shape).to(dtype=torch.float32)
-        scale = torch.min(self_min_size / min_size, self_max_size / max_size)
+        if torch.jit.is_scripting() or torchvision._is_tracing():
+            min_size = torch.min(im_shape).to(dtype=torch.float32)
+            max_size = torch.max(im_shape).to(dtype=torch.float32)
+            self_min_size_f = float(self_min_size)
+            self_max_size_f = float(self_max_size)
+            scale = torch.min(self_min_size_f / min_size, self_max_size_f / max_size)
+
+            if torchvision._is_tracing():
+                scale_factor = _fake_cast_onnx(scale)
+            else:
+                scale_factor = scale.item()
 
-        if torchvision._is_tracing():
-            scale_factor = _fake_cast_onnx(scale)
         else:
-            scale_factor = scale.item()
+            # Do it the normal way
+            min_size = min(im_shape)
+            max_size = max(im_shape)
+            scale_factor = min(self_min_size / min_size, self_max_size / max_size)
+
         recompute_scale_factor = True
 
     image = torch.nn.functional.interpolate(
@@ -159,8 +169,7 @@ class GeneralizedRCNNTransform(nn.Module):
     def torch_choice(self, k: List[int]) -> int:
         """
         Implements `random.choice` via torch ops, so it can be compiled with
-        TorchScript. Remove if https://github.com/pytorch/pytorch/issues/25803
-        is fixed.
+        TorchScript and we use PyTorch's RNG (not native RNG)
         """
         index = int(torch.empty(1).uniform_(0.0, float(len(k))).item())
         return k[index]
@@ -174,11 +183,10 @@ class GeneralizedRCNNTransform(nn.Module):
         if self.training:
             if self._skip_resize:
                 return image, target
-            size = float(self.torch_choice(self.min_size))
+            size = self.torch_choice(self.min_size)
         else:
-            # FIXME assume for now that testing uses the largest scale
-            size = float(self.min_size[-1])
-        image, target = _resize_image_and_masks(image, size, float(self.max_size), target, self.fixed_size)
+            size = self.min_size[-1]
+        image, target = _resize_image_and_masks(image, size, self.max_size, target, self.fixed_size)
 
         if target is None:
             return image, target