Fixed floor_divide deprecation warnings seen in pytest output (#3672)

test/test_datasets_samplers.py

@@ -45,7 +45,7 @@ class Tester(unittest.TestCase):
             sampler = RandomClipSampler(video_clips, 3)
             self.assertEqual(len(sampler), 3 * 3)
             indices = torch.tensor(list(iter(sampler)))
-            videos = indices // 5
+            videos = torch.div(indices, 5, rounding_mode='floor')
             v_idxs, count = torch.unique(videos, return_counts=True)
             self.assertTrue(v_idxs.equal(torch.tensor([0, 1, 2])))
             self.assertTrue(count.equal(torch.tensor([3, 3, 3])))
@@ -62,7 +62,7 @@ class Tester(unittest.TestCase):
             indices.remove(0)
             indices.remove(1)
             indices = torch.tensor(indices) - 2
-            videos = indices // 5
+            videos = torch.div(indices, 5, rounding_mode='floor')
             v_idxs, count = torch.unique(videos, return_counts=True)
             self.assertTrue(v_idxs.equal(torch.tensor([0, 1])))
             self.assertTrue(count.equal(torch.tensor([3, 3])))
@@ -73,7 +73,7 @@ class Tester(unittest.TestCase):
             sampler = UniformClipSampler(video_clips, 3)
             self.assertEqual(len(sampler), 3 * 3)
             indices = torch.tensor(list(iter(sampler)))
-            videos = indices // 5
+            videos = torch.div(indices, 5, rounding_mode='floor')
             v_idxs, count = torch.unique(videos, return_counts=True)
             self.assertTrue(v_idxs.equal(torch.tensor([0, 1, 2])))
             self.assertTrue(count.equal(torch.tensor([3, 3, 3])))

torchvision/datasets/celeba.py

@@ -104,7 +104,8 @@ class CelebA(VisionDataset):
         self.bbox = bbox.data[mask]
         self.landmarks_align = landmarks_align.data[mask]
         self.attr = attr.data[mask]
-        self.attr = (self.attr + 1) // 2  # map from {-1, 1} to {0, 1}
+        # map from {-1, 1} to {0, 1}
+        self.attr = torch.div(self.attr + 1, 2, rounding_mode='floor')
         self.attr_names = attr.header
 
     def _load_csv(

torchvision/models/detection/retinanet.py

@@ -428,7 +428,7 @@ class RetinaNet(nn.Module):
                 scores_per_level, idxs = scores_per_level.topk(num_topk)
                 topk_idxs = topk_idxs[idxs]
 
-                anchor_idxs = topk_idxs // num_classes
+                anchor_idxs = torch.div(topk_idxs, num_classes, rounding_mode='floor')
                 labels_per_level = topk_idxs % num_classes
 
                 boxes_per_level = self.box_coder.decode_single(box_regression_per_level[anchor_idxs],

torchvision/models/detection/roi_heads.py

@@ -266,7 +266,7 @@ def heatmaps_to_keypoints(maps, rois):
         pos = roi_map.reshape(num_keypoints, -1).argmax(dim=1)
 
         x_int = pos % w
-        y_int = (pos - x_int) // w
+        y_int = torch.div(pos - x_int, w, rounding_mode='floor')
         # assert (roi_map_probs[k, y_int, x_int] ==
         #         roi_map_probs[k, :, :].max())
         x = (x_int.float() + 0.5) * width_correction

torchvision/transforms/functional_tensor.py

@@ -97,7 +97,7 @@ def convert_image_dtype(image: torch.Tensor, dtype: torch.dtype = torch.float) -
             # factor should be forced to int for torch jit script
             # otherwise factor is a float and image // factor can produce different results
             factor = int((input_max + 1) // (output_max + 1))
-            image = image // factor
+            image = torch.div(image, factor, rounding_mode='floor')
             return image.to(dtype)
         else:
             # factor should be forced to int for torch jit script
@@ -908,11 +908,13 @@ def _scale_channel(img_chan):
         hist = torch.bincount(img_chan.view(-1), minlength=256)
 
     nonzero_hist = hist[hist != 0]
-    step = nonzero_hist[:-1].sum() // 255
+    step = torch.div(nonzero_hist[:-1].sum(), 255, rounding_mode='floor')
     if step == 0:
         return img_chan
 
-    lut = (torch.cumsum(hist, 0) + (step // 2)) // step
+    lut = torch.div(
+        torch.cumsum(hist, 0) + torch.div(step, 2, rounding_mode='floor'),
+        step, rounding_mode='floor')
     lut = torch.nn.functional.pad(lut, [1, 0])[:-1].clamp(0, 255)
 
     return lut[img_chan.to(torch.int64)].to(torch.uint8)