feat: add loader to Omniglot and INaturalist.

Author: GdoongMathew

test/test_datasets.py

@@ -1146,6 +1146,7 @@ class OmniglotTestCase(datasets_utils.ImageDatasetTestCase):
     DATASET_CLASS = datasets.Omniglot
 
     ADDITIONAL_CONFIGS = combinations_grid(background=(True, False))
+    SUPPORT_TV_IMAGE_DECODE = True
 
     def inject_fake_data(self, tmpdir, config):
         target_folder = (
@@ -1910,6 +1911,7 @@ class INaturalistTestCase(datasets_utils.ImageDatasetTestCase):
         target_type=("kingdom", "full", "genus", ["kingdom", "phylum", "class", "order", "family", "genus", "full"]),
         version=("2021_train",),
     )
+    SUPPORT_TV_IMAGE_DECODE = True
 
     def inject_fake_data(self, tmpdir, config):
         categories = [

torchvision/datasets/inaturalist.py

@@ -62,6 +62,9 @@ class INaturalist(VisionDataset):
         download (bool, optional): If true, downloads the dataset from the internet and
             puts it in root directory. If dataset is already downloaded, it is not
             downloaded again.
+        loader (callable, optional): A function to load an image given its path.
+            By default, it uses PIL as its image loader, but users could also pass in
+            ``torchvision.io.decode_image`` for decoding image data into tensors directly.
     """
 
     def __init__(
@@ -72,6 +75,7 @@ def __init__(
         transform: Optional[Callable] = None,
         target_transform: Optional[Callable] = None,
         download: bool = False,
+        loader: Optional[Callable[[Union[str, Path]], Any]] = None,
     ) -> None:
         self.version = verify_str_arg(version, "version", DATASET_URLS.keys())
 
@@ -109,6 +113,8 @@ def __init__(
             for fname in files:
                 self.index.append((dir_index, fname))
 
+        self.loader = loader or Image.open
+
     def _init_2021(self) -> None:
         """Initialize based on 2021 layout"""
 
@@ -178,7 +184,7 @@ def __getitem__(self, index: int) -> Tuple[Any, Any]:
         """
 
         cat_id, fname = self.index[index]
-        img = Image.open(os.path.join(self.root, self.all_categories[cat_id], fname))
+        img = self.loader(os.path.join(self.root, self.all_categories[cat_id], fname))
 
         target: Any = []
         for t in self.target_type:

torchvision/datasets/omniglot.py

@@ -23,6 +23,9 @@ class Omniglot(VisionDataset):
         download (bool, optional): If true, downloads the dataset zip files from the internet and
             puts it in root directory. If the zip files are already downloaded, they are not
             downloaded again.
+        loader (callable, optional): A function to load an image given its path.
+            By default, it uses PIL as its image loader, but users could also pass in
+            ``torchvision.io.decode_image`` for decoding image data into tensors directly.
     """
 
     folder = "omniglot-py"
@@ -39,6 +42,7 @@ def __init__(
         transform: Optional[Callable] = None,
         target_transform: Optional[Callable] = None,
         download: bool = False,
+        loader: Optional[Callable[[Union[str, Path]], Any]] = None,
     ) -> None:
         super().__init__(join(root, self.folder), transform=transform, target_transform=target_transform)
         self.background = background
@@ -59,6 +63,7 @@ def __init__(
             for idx, character in enumerate(self._characters)
         ]
         self._flat_character_images: List[Tuple[str, int]] = sum(self._character_images, [])
+        self.loader = loader
 
     def __len__(self) -> int:
         return len(self._flat_character_images)
@@ -73,7 +78,7 @@ def __getitem__(self, index: int) -> Tuple[Any, Any]:
         """
         image_name, character_class = self._flat_character_images[index]
         image_path = join(self.target_folder, self._characters[character_class], image_name)
-        image = Image.open(image_path, mode="r").convert("L")
+        image = Image.open(image_path, mode="r").convert("L") if self.loader is None else self.loader(image_path)
 
         if self.transform:
             image = self.transform(image)