Fix typos and grammar errors (#7065)

* fix typos throughout the code base

* fix grammar

* revert formatting changes to gallery

* revert 'an uXX'

* remove 'number of the best'

test/test_ops.py

@@ -87,7 +87,7 @@ class RoIOpTester(ABC):
         x_dtype = self.dtype if x_dtype is None else x_dtype
         rois_dtype = self.dtype if rois_dtype is None else rois_dtype
         pool_size = 5
-        # n_channels % (pool_size ** 2) == 0 required for PS opeartions.
+        # n_channels % (pool_size ** 2) == 0 required for PS operations.
         n_channels = 2 * (pool_size**2)
         x = torch.rand(2, n_channels, 10, 10, dtype=x_dtype, device=device)
         if not contiguous:
@@ -647,11 +647,11 @@ class TestNMS:
     @pytest.mark.parametrize("scale, zero_point", ((1, 0), (2, 50), (3, 10)))
     def test_qnms(self, iou, scale, zero_point):
         # Note: we compare qnms vs nms instead of qnms vs reference implementation.
-        # This is because with the int convertion, the trick used in _create_tensors_with_iou
+        # This is because with the int conversion, the trick used in _create_tensors_with_iou
         # doesn't really work (in fact, nms vs reference implem will also fail with ints)
         err_msg = "NMS and QNMS give different results for IoU={}"
         boxes, scores = self._create_tensors_with_iou(1000, iou)
-        scores *= 100  # otherwise most scores would be 0 or 1 after int convertion
+        scores *= 100  # otherwise most scores would be 0 or 1 after int conversion
 
         qboxes = torch.quantize_per_tensor(boxes, scale=scale, zero_point=zero_point, dtype=torch.quint8)
         qscores = torch.quantize_per_tensor(scores, scale=scale, zero_point=zero_point, dtype=torch.quint8)

test/test_prototype_transforms.py

@@ -135,7 +135,7 @@ class TestSmoke:
     def test_mixup_cutmix(self, transform, input):
         transform(input)
 
-        # add other data that should bypass and wont raise any error
+        # add other data that should bypass and won't raise any error
         input_copy = dict(input)
         input_copy["path"] = "/path/to/somewhere"
         input_copy["num"] = 1234

test/test_transforms.py

@@ -1818,7 +1818,7 @@ def test_random_erasing(seed):
     tol = 0.05
     assert 1 / 3 - tol <= aspect_ratio <= 3 + tol
 
-    # Make sure that h > w and h < w are equaly likely (log-scale sampling)
+    # Make sure that h > w and h < w are equally likely (log-scale sampling)
     aspect_ratios = []
     random.seed(42)
     trial = 1000

test/test_utils.py

@@ -184,7 +184,7 @@ def test_draw_no_boxes():
     boxes = torch.full((0, 4), 0, dtype=torch.float)
     with pytest.warns(UserWarning, match=re.escape("boxes doesn't contain any box. No box was drawn")):
         res = utils.draw_bounding_boxes(img, boxes)
-        # Check that the function didnt change the image
+        # Check that the function didn't change the image
         assert res.eq(img).all()
 
 
@@ -209,7 +209,7 @@ def test_draw_segmentation_masks(colors, alpha):
 
     # For testing we enforce that there's no overlap between the masks. The
     # current behaviour is that the last mask's color will take priority when
-    # masks overlap, but this makes testing slightly harder so we don't really
+    # masks overlap, but this makes testing slightly harder, so we don't really
     # care
     overlap = masks[0] & masks[1]
     masks[:, overlap] = False
@@ -283,7 +283,7 @@ def test_draw_no_segmention_mask():
     masks = torch.full((0, 100, 100), 0, dtype=torch.bool)
     with pytest.warns(UserWarning, match=re.escape("masks doesn't contain any mask. No mask was drawn")):
         res = utils.draw_segmentation_masks(img, masks)
-        # Check that the function didnt change the image
+        # Check that the function didn't change the image
         assert res.eq(img).all()
 
 

test/test_video_reader.py

@@ -127,7 +127,7 @@ def _read_from_stream(container, start_pts, end_pts, stream, stream_name, buffer
             ascending order. We need to decode more frames even when we meet end
             pts
     """
-    # seeking in the stream is imprecise. Thus, seek to an ealier PTS by a margin
+    # seeking in the stream is imprecise. Thus, seek to an earlier PTS by a margin
     margin = 1
     seek_offset = max(start_pts - margin, 0)
 

torchvision/csrc/io/decoder/defs.h

@@ -301,7 +301,7 @@ struct DecoderMetadata {
 };
 /**
  * Abstract class for decoding media bytes
- * It has two diffrent modes. Internal media bytes retrieval for given uri and
+ * It has two different modes. Internal media bytes retrieval for given uri and
  * external media bytes provider in case of memory streams
  */
 class MediaDecoder {

torchvision/csrc/io/decoder/memory_buffer.cpp

@@ -61,7 +61,7 @@ DecoderInCallback MemoryBuffer::getCallback(
         }
         // seek mode
         if (!timeoutMs) {
-          // seek capabilty, yes - supported
+          // seek capability, yes - supported
           return 0;
         }
         return object.seek(size, whence);

torchvision/csrc/io/decoder/sync_decoder_test.cpp

@@ -368,7 +368,7 @@ TEST(SyncDecoder, TestMemoryBufferNoSeekableWithFullRead) {
         }
         // seek mode
         if (!timeoutMs) {
-          // seek capabilty, yes - no
+          // seek capability, yes - no
           return -1;
         }
         return object.seek(size, whence);
@@ -408,7 +408,7 @@ TEST(SyncDecoder, TestMemoryBufferNoSeekableWithPartialRead) {
         }
         // seek mode
         if (!timeoutMs) {
-          // seek capabilty, yes - no
+          // seek capability, yes - no
           return -1;
         }
         return object.seek(size, whence);

torchvision/csrc/io/video/video.cpp

@@ -319,14 +319,14 @@ void Video::Seek(double ts, bool fastSeek = false) {
 std::tuple<torch::Tensor, double> Video::Next() {
   TORCH_CHECK(initialized, "Video object has to be initialized first");
   // if failing to decode simply return a null tensor (note, should we
-  // raise an exeption?)
+  // raise an exception?)
   double frame_pts_s;
   torch::Tensor outFrame = torch::zeros({0}, torch::kByte);
 
   // decode single frame
   DecoderOutputMessage out;
   int64_t res = decoder.decode(&out, decoderTimeoutMs);
-  // if successfull
+  // if successful
   if (res == 0) {
     frame_pts_s = double(double(out.header.pts) * 1e-6);
 

torchvision/csrc/io/video/video.h

@@ -42,8 +42,8 @@ struct Video : torch::CustomClassHolder {
  private:
   bool succeeded = false; // decoder init flag
   // seekTS and doSeek act as a flag - if it's not set, next function simply
-  // retruns the next frame. If it's set, we look at the global seek
-  // time in comination with any_frame settings
+  // returns the next frame. If it's set, we look at the global seek
+  // time in combination with any_frame settings
   double seekTS = -1;
 
   bool initialized = false;

torchvision/csrc/ops/cpu/roi_align_kernel.cpp

@@ -60,7 +60,7 @@ void roi_align_forward_kernel_impl(
     // When the grid is empty, output zeros.
     const T count = std::max(roi_bin_grid_h * roi_bin_grid_w, 1); // e.g. = 4
 
-    // we want to precalculate indices and weights shared by all chanels,
+    // we want to precalculate indices and weights shared by all channels,
     // this is the key point of optimization
     std::vector<detail::PreCalc<T>> pre_calc(
         roi_bin_grid_h * roi_bin_grid_w * pooled_width * pooled_height);

torchvision/csrc/ops/quantized/cpu/qroi_align_kernel.cpp

@@ -164,7 +164,7 @@ void qroi_align_forward_kernel_impl(
     const float count =
         std::max(roi_bin_grid_h * roi_bin_grid_w, 1); // e.g. = 4
 
-    // we want to precalculate indices and weights shared by all chanels,
+    // we want to precalculate indices and weights shared by all channels,
     // this is the key point of optimization
     std::vector<detail::PreCalc<float>> pre_calc(
         roi_bin_grid_h * roi_bin_grid_w * pooled_width * pooled_height);

torchvision/datasets/_stereo_matching.py

@@ -424,9 +424,9 @@ class Middlebury2014Stereo(StereoMatchingDataset):
         split (string, optional): The dataset split of scenes, either "train" (default), "test", or "additional"
         use_ambient_views (boolean, optional): Whether to use different expose or lightning views when possible.
             The dataset samples with equal probability between ``[im1.png, im1E.png, im1L.png]``.
-        calibration (string, optional): Wether or not to use the calibrated (default) or uncalibrated scenes.
+        calibration (string, optional): Whether or not to use the calibrated (default) or uncalibrated scenes.
         transforms (callable, optional): A function/transform that takes in a sample and returns a transformed version.
-        download (boolean, optional): Wether or not to download the dataset in the ``root`` directory.
+        download (boolean, optional): Whether or not to download the dataset in the ``root`` directory.
     """
 
     splits = {
@@ -720,7 +720,7 @@ class CREStereo(StereoMatchingDataset):
 class FallingThingsStereo(StereoMatchingDataset):
     """`FallingThings <https://research.nvidia.com/publication/2018-06_falling-things-synthetic-dataset-3d-object-detection-and-pose-estimation>`_ dataset.
 
-    The dataset is expected to have the following structre: ::
+    The dataset is expected to have the following structure: ::
 
         root
             FallingThings
@@ -825,7 +825,7 @@ class SceneFlowStereo(StereoMatchingDataset):
     """Dataset interface for `Scene Flow <https://lmb.informatik.uni-freiburg.de/resources/datasets/SceneFlowDatasets.en.html>`_ datasets.
     This interface provides access to the `FlyingThings3D, `Monkaa` and `Driving` datasets.
 
-    The dataset is expected to have the following structre: ::
+    The dataset is expected to have the following structure: ::
 
         root
             SceneFlow
@@ -1031,7 +1031,7 @@ class SintelStereo(StereoMatchingDataset):
         disparity_map = r * 4 + g / (2**6) + b / (2**14)
         # reshape into (C, H, W) format
         disparity_map = np.transpose(disparity_map, (2, 0, 1))
-        # find the appropiate file paths
+        # find the appropriate file paths
         occlued_mask_path, out_of_frame_mask_path = self._get_occlussion_mask_paths(file_path)
         # occlusion masks
         valid_mask = np.asarray(Image.open(occlued_mask_path)) == 0
@@ -1058,7 +1058,7 @@ class SintelStereo(StereoMatchingDataset):
 class InStereo2k(StereoMatchingDataset):
     """`InStereo2k <https://github.com/YuhuaXu/StereoDataset>`_ dataset.
 
-    The dataset is expected to have the following structre: ::
+    The dataset is expected to have the following structure: ::
 
         root
             InStereo2k

torchvision/datasets/celeba.py

@@ -41,7 +41,7 @@ class CelebA(VisionDataset):
     """
 
     base_folder = "celeba"
-    # There currently does not appear to be a easy way to extract 7z in python (without introducing additional
+    # There currently does not appear to be an easy way to extract 7z in python (without introducing additional
     # dependencies). The "in-the-wild" (not aligned+cropped) images are only in 7z, so they are not available
     # right now.
     file_list = [

torchvision/datasets/cityscapes.py

@@ -177,7 +177,7 @@ class Cityscapes(VisionDataset):
             index (int): Index
         Returns:
             tuple: (image, target) where target is a tuple of all target types if target_type is a list with more
-            than one item. Otherwise target is a json object if target_type="polygon", else the image segmentation.
+            than one item. Otherwise, target is a json object if target_type="polygon", else the image segmentation.
         """
 
         image = Image.open(self.images[index]).convert("RGB")

torchvision/datasets/country211.py

@@ -11,7 +11,7 @@ class Country211(ImageFolder):
     This dataset was built by filtering the images from the YFCC100m dataset
     that have GPS coordinate corresponding to a ISO-3166 country code. The
     dataset is balanced by sampling 150 train images, 50 validation images, and
-    100 test images images for each country.
+    100 test images for each country.
 
     Args:
         root (string): Root directory of the dataset.

torchvision/datasets/hmdb51.py

@@ -102,7 +102,7 @@ class HMDB51(VisionDataset):
             output_format=output_format,
         )
         # we bookkeep the full version of video clips because we want to be able
-        # to return the meta data of full version rather than the subset version of
+        # to return the metadata of full version rather than the subset version of
         # video clips
         self.full_video_clips = video_clips
         self.fold = fold

torchvision/datasets/mnist.py

@@ -366,7 +366,7 @@ class QMNIST(MNIST):
             that takes in the target and transforms it.
         train (bool,optional,compatibility): When argument 'what' is
             not specified, this boolean decides whether to load the
-            training set ot the testing set.  Default: True.
+            training set or the testing set.  Default: True.
     """
 
     subsets = {"train": "train", "test": "test", "test10k": "test", "test50k": "test", "nist": "nist"}

torchvision/datasets/places365.py

@@ -15,7 +15,7 @@ class Places365(VisionDataset):
         root (string): Root directory of the Places365 dataset.
         split (string, optional): The dataset split. Can be one of ``train-standard`` (default), ``train-challenge``,
             ``val``.
-        small (bool, optional): If ``True``, uses the small images, i. e. resized to 256 x 256 pixels, instead of the
+        small (bool, optional): If ``True``, uses the small images, i.e. resized to 256 x 256 pixels, instead of the
             high resolution ones.
         download (bool, optional): If ``True``, downloads the dataset components and places them in ``root``. Already
             downloaded archives are not downloaded again.
@@ -32,7 +32,7 @@ class Places365(VisionDataset):
         targets (list): The class_index value for each image in the dataset
 
     Raises:
-        RuntimeError: If ``download is False`` and the meta files, i. e. the devkit, are not present or corrupted.
+        RuntimeError: If ``download is False`` and the meta files, i.e. the devkit, are not present or corrupted.
         RuntimeError: If ``download is True`` and the image archive is already extracted.
     """
     _SPLITS = ("train-standard", "train-challenge", "val")

torchvision/datasets/stl10.py

@@ -15,7 +15,7 @@ class STL10(VisionDataset):
         root (string): Root directory of dataset where directory
             ``stl10_binary`` exists.
         split (string): One of {'train', 'test', 'unlabeled', 'train+unlabeled'}.
-            Accordingly dataset is selected.
+            Accordingly, dataset is selected.
         folds (int, optional): One of {0-9} or None.
             For training, loads one of the 10 pre-defined folds of 1k samples for the
             standard evaluation procedure. If no value is passed, loads the 5k samples.