From efcf0cb5798e80c8ec1aebc1f70634371de8b8d3 Mon Sep 17 00:00:00 2001
From: Abhijit Deo <72816663+abhi-glitchhg@users.noreply.github.com>
Date: Thu, 6 Jan 2022 13:19:55 +0530
Subject: [PATCH 01/15] [WIP] *added stanford_cars

---
 torchvision/datasets/standford_cars.py | 129 +++++++++++++++++++++++++
 1 file changed, 129 insertions(+)
 create mode 100644 torchvision/datasets/standford_cars.py

diff --git a/torchvision/datasets/standford_cars.py b/torchvision/datasets/standford_cars.py
new file mode 100644
index 00000000000..5848bd569f1
--- /dev/null
+++ b/torchvision/datasets/standford_cars.py
@@ -0,0 +1,129 @@
+import os
+import os.path
+from typing import Callable, Optional
+
+import scipy.io as io
+from PIL import Image
+
+from .utils import download_and_extract_archive, verify_str_arg, download_url
+from .vision import VisionDataset
+
+
+class StanfordCars(VisionDataset):
+    """`Stanford Cars <https://ai.stanford.edu/~jkrause/cars/car_dataset.html>`_ Dataset
+
+    .. warning::
+
+        This class needs `scipy <https://docs.scipy.org/doc/>`_ to load target files from `.mat` format.
+
+    Args:
+        root (string): Root directory of dataset
+        train (bool, optional):
+        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.
+        transform (callable, optional): A function/transform that  takes in an PIL image
+            and returns a transformed version. E.g, ``transforms.RandomCrop``
+        target_transform (callable, optional): A function/transform that takes in the
+            target and transforms it."""
+
+    urls = (
+        "https://ai.stanford.edu/~jkrause/car196/cars_test.tgz",
+        "https://ai.stanford.edu/~jkrause/car196/cars_train.tgz",
+    )
+
+    md5s = ("4ce7ebf6a94d07f1952d94dd34c4d501", "065e5b463ae28d29e77c1b4b166cfe61")
+
+    annot_urls = (
+        "https://ai.stanford.edu/~jkrause/car196/cars_test_annos_withlabels.mat",
+        "https://ai.stanford.edu/~jkrause/cars/car_devkit.tgz",
+    )
+    extension = ".jpg"
+
+    def __init__(
+        self,
+        root: str,
+        train: bool = True,
+        transform: Optional[Callable] = None,
+        target_transform: Optional[Callable] = None,
+        download: bool = False,
+    ) -> None:
+        super().__init__(root, transform=transform, target_transform=target_transform)
+        self._download = download
+        self.root = os.path.expanduser(root)
+        self.train = train
+        self.url = self.urls[self.train]
+
+        if self._download:
+            self.download()
+
+        if not self._check_exists():
+            raise RuntimeError("Dataset not found. You can use download=True to download it")
+
+        self.class_names = self._get_class_names()
+        self.annotations = self._get_annotations()
+
+    def _get_class_names(self):
+        """
+        Returns Mapping of class ids to class names in form of Dictionary
+        """
+        meta_data = io.loadmat(os.path.join(self.root, "devkit/cars_meta.mat"))
+        class_names = meta_data["class_names"][0]
+        return {
+            # Format class names appropriately for directory creation.
+            class_name[0].replace(" ", "_").replace("/", "_"): i
+            for i, class_name in enumerate(class_names)
+        }
+
+    def _get_annotations(self):
+        """
+        Returns Annotations for training data and testing data
+        """
+        annotations = None
+        if self.train:
+            annotations = io.loadmat(os.path.join(self.root, "devkit/cars_train_annos.mat"))
+        else:
+            annotations = io.loadmat(os.path.join(self.root, "cars_test_annos_withlabels.mat"))
+        return annotations["annotations"][0]
+
+    def __len__(self):
+        return self.annotations.shape[0]
+
+    def __getitem__(self, idx: int):
+        """Returns pil_image and class_id for given index"""
+        target_id = self.annotations[idx][4][0, 0]
+        image_name = self.annotations[idx][5][0]
+        # 0,1,2,3 correspond to bounding boxes
+
+        image_path = os.path.join(self.root, f"cars_{'train' if self.train else 'test' }", image_name)
+        # Beware: Stanford cars targets starts at 1
+        target_id = target_id - 1
+        pil_image = Image.open(image_path).convert("RGB")
+
+        if self.transform is not None:
+            pil_image = self.transform(pil_image)
+        if self.target_transform is not None:
+            pil_image = self.target_transform(pil_image)
+        return pil_image, target_id
+
+
+    def download(self) -> None:
+        if self._check_exists():
+            return
+        else:
+            download_and_extract_archive(
+                url=self.urls[self.train],
+                download_root=self.root,
+                extract_root=self.root,
+                md5=self.md5s[self.train]
+            )
+            download_and_extract_archive(
+                    url=self.annot_urls[1], download_root=self.root, extract_root=self.root ,md5="c3b158d763b6e2245038c8ad08e45376"
+                )
+            if not self.train:
+                download_url(
+                    url=self.annot_urls[0], filename="cars_test_annos_withlabels.mat", root=self.root ,md5="b0a2b23655a3edd16d84508592a98d10"
+                )
+    def _check_exists(self) -> bool:
+        return os.path.exists(os.path.join(self.root, f"cars_{'train' if self.train else 'test'}")) and os.path.isdir(os.path.join(self.root, f"cars_{'train' if self.train else 'test'}")) and os.path.exists(os.path.join(self.root,"devkit/cars_meta.mat"))
+

From 62ca7bf9950589fb5be1f004764f5522e108ec63 Mon Sep 17 00:00:00 2001
From: Abhijit Deo <72816663+abhi-glitchhg@users.noreply.github.com>
Date: Thu, 6 Jan 2022 13:23:07 +0530
Subject: [PATCH 02/15] [WIP] added stanfordCars to docs

---
 docs/source/datasets.rst         | 1 +
 torchvision/datasets/__init__.py | 2 ++
 2 files changed, 3 insertions(+)

diff --git a/docs/source/datasets.rst b/docs/source/datasets.rst
index 8a71142f9e7..5b903aa3873 100644
--- a/docs/source/datasets.rst
+++ b/docs/source/datasets.rst
@@ -68,6 +68,7 @@ You can also create your own datasets using the provided :ref:`base classes <bas
     SBU
     SEMEION
     Sintel
+    StanfordCars
     STL10
     SVHN
     UCF101
diff --git a/torchvision/datasets/__init__.py b/torchvision/datasets/__init__.py
index a5533cc6003..643b1d60f64 100644
--- a/torchvision/datasets/__init__.py
+++ b/torchvision/datasets/__init__.py
@@ -24,6 +24,7 @@
 from .sbd import SBDataset
 from .sbu import SBU
 from .semeion import SEMEION
+from .standford_cars import StanfordCars
 from .stl10 import STL10
 from .svhn import SVHN
 from .ucf101 import UCF101
@@ -47,6 +48,7 @@
     "QMNIST",
     "MNIST",
     "KMNIST",
+    "StanfordCars",
     "STL10",
     "SVHN",
     "PhotoTour",

From 4b123e4f9d10b4a747ca579048f8812b5f603076 Mon Sep 17 00:00:00 2001
From: Abhijit Deo <72816663+abhi-glitchhg@users.noreply.github.com>
Date: Thu, 6 Jan 2022 13:41:57 +0530
Subject: [PATCH 03/15] [WIP] minor edits

---
 torchvision/datasets/standford_cars.py | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/torchvision/datasets/standford_cars.py b/torchvision/datasets/standford_cars.py
index 5848bd569f1..77156e06c7f 100644
--- a/torchvision/datasets/standford_cars.py
+++ b/torchvision/datasets/standford_cars.py
@@ -63,7 +63,7 @@ def __init__(
         self.class_names = self._get_class_names()
         self.annotations = self._get_annotations()
 
-    def _get_class_names(self):
+    def _get_class_names(self) -> dict:
         """
         Returns Mapping of class ids to class names in form of Dictionary
         """
@@ -86,10 +86,10 @@ def _get_annotations(self):
             annotations = io.loadmat(os.path.join(self.root, "cars_test_annos_withlabels.mat"))
         return annotations["annotations"][0]
 
-    def __len__(self):
+    def __len__(self) -> int:
         return self.annotations.shape[0]
 
-    def __getitem__(self, idx: int):
+    def __getitem__(self, idx: int) -> (Image,int):
         """Returns pil_image and class_id for given index"""
         target_id = self.annotations[idx][4][0, 0]
         image_name = self.annotations[idx][5][0]

From 52cd5b9a9a9d3d2f7ec66d43e2e0f7254f7bdeea Mon Sep 17 00:00:00 2001
From: Abhijit Deo <72816663+abhi-glitchhg@users.noreply.github.com>
Date: Thu, 6 Jan 2022 14:07:05 +0530
Subject: [PATCH 04/15] [WIP] minor edits

---
 torchvision/datasets/standford_cars.py | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

diff --git a/torchvision/datasets/standford_cars.py b/torchvision/datasets/standford_cars.py
index 77156e06c7f..17bcd00da98 100644
--- a/torchvision/datasets/standford_cars.py
+++ b/torchvision/datasets/standford_cars.py
@@ -5,7 +5,7 @@
 import scipy.io as io
 from PIL import Image
 
-from .utils import download_and_extract_archive, verify_str_arg, download_url
+from .utils import download_and_extract_archive, download_url
 from .vision import VisionDataset
 
 
@@ -125,5 +125,4 @@ def download(self) -> None:
                     url=self.annot_urls[0], filename="cars_test_annos_withlabels.mat", root=self.root ,md5="b0a2b23655a3edd16d84508592a98d10"
                 )
     def _check_exists(self) -> bool:
-        return os.path.exists(os.path.join(self.root, f"cars_{'train' if self.train else 'test'}")) and os.path.isdir(os.path.join(self.root, f"cars_{'train' if self.train else 'test'}")) and os.path.exists(os.path.join(self.root,"devkit/cars_meta.mat"))
-
+        return os.path.exists(os.path.join(self.root, f"cars_{'train' if self.train else 'test'}")) and os.path.isdir(os.path.join(self.root, f"cars_{'train' if self.train else 'test'}")) and os.path.exists(os.path.join(self.root,"devkit/cars_meta.mat"))
\ No newline at end of file

From 346036eb462bacb84476eafc8ba36e6733c9b3bd Mon Sep 17 00:00:00 2001
From: Abhijit Deo <72816663+abhi-glitchhg@users.noreply.github.com>
Date: Sat, 8 Jan 2022 10:59:10 +0530
Subject: [PATCH 05/15] edited StanfordCars class

---
 test/test_datasets.py                  |  7 ++
 torchvision/datasets/standford_cars.py | 91 +++++++++++++++-----------
 2 files changed, 59 insertions(+), 39 deletions(-)

diff --git a/test/test_datasets.py b/test/test_datasets.py
index 5b5056e7ebe..c89eb1e8262 100644
--- a/test/test_datasets.py
+++ b/test/test_datasets.py
@@ -2274,6 +2274,13 @@ def inject_fake_data(self, tmpdir, config):
 
         return num_samples
 
+class StanfordCarsTestCase(datasets_utils.ImageDatasetTestCase):
+    DATASET_CLASS = datasets.StanfordCars
+    REQUIRED_PACKAGES = ("scipy",)
+    FEATURE_TYPES = (PIL.Image.Image, int)
+    ADDITIONAL_CONFIGS = datasets_utils.combinations_grid(train=(True,False))
+
+
 
 if __name__ == "__main__":
     unittest.main()
diff --git a/torchvision/datasets/standford_cars.py b/torchvision/datasets/standford_cars.py
index 17bcd00da98..3d13e5b07b3 100644
--- a/torchvision/datasets/standford_cars.py
+++ b/torchvision/datasets/standford_cars.py
@@ -1,8 +1,6 @@
 import os
 import os.path
 from typing import Callable, Optional
-
-import scipy.io as io
 from PIL import Image
 
 from .utils import download_and_extract_archive, download_url
@@ -19,85 +17,96 @@ class StanfordCars(VisionDataset):
     Args:
         root (string): Root directory of dataset
         train (bool, optional):
-        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.
         transform (callable, optional): A function/transform that  takes in an PIL image
             and returns a transformed version. E.g, ``transforms.RandomCrop``
         target_transform (callable, optional): A function/transform that takes in the
-            target and transforms it."""
+            target and transforms it.
+        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."""
 
     urls = (
         "https://ai.stanford.edu/~jkrause/car196/cars_test.tgz",
         "https://ai.stanford.edu/~jkrause/car196/cars_train.tgz",
     )
 
-    md5s = ("4ce7ebf6a94d07f1952d94dd34c4d501", "065e5b463ae28d29e77c1b4b166cfe61")
+    md5s = ("4ce7ebf6a94d07f1952d94dd34c4d501", "065e5b463ae28d29e77c1b4b166cfe61")  # md5checksum for test and train
 
     annot_urls = (
         "https://ai.stanford.edu/~jkrause/car196/cars_test_annos_withlabels.mat",
         "https://ai.stanford.edu/~jkrause/cars/car_devkit.tgz",
     )
-    extension = ".jpg"
+    annot_md5s = (
+        "b0a2b23655a3edd16d84508592a98d10",
+        "c3b158d763b6e2245038c8ad08e45376",
+    )
 
     def __init__(
-        self,
-        root: str,
-        train: bool = True,
-        transform: Optional[Callable] = None,
-        target_transform: Optional[Callable] = None,
-        download: bool = False,
+            self,
+            root: str,
+            train: bool = True,
+            transform: Optional[Callable] = None,
+            target_transform: Optional[Callable] = None,
+            download: bool = False,
     ) -> None:
+
+        try:
+            from scipy.io import loadmat
+
+            self._loadmat = loadmat
+        except ImportError:
+            raise RuntimeError("Scipy is not found. This dataset needs to have scipy installed: pip install scipy")
+
         super().__init__(root, transform=transform, target_transform=target_transform)
-        self._download = download
-        self.root = os.path.expanduser(root)
+
         self.train = train
-        self.url = self.urls[self.train]
 
-        if self._download:
+        if download:
             self.download()
 
         if not self._check_exists():
             raise RuntimeError("Dataset not found. You can use download=True to download it")
 
-        self.class_names = self._get_class_names()
-        self.annotations = self._get_annotations()
+        self._samples = self._make_dataset()
+        self.classes = self._get_class_names()  # class_id to class_name mapping
 
     def _get_class_names(self) -> dict:
         """
         Returns Mapping of class ids to class names in form of Dictionary
         """
-        meta_data = io.loadmat(os.path.join(self.root, "devkit/cars_meta.mat"))
+        meta_data = self._loadmat(os.path.join(self.root, "devkit/cars_meta.mat"))
         class_names = meta_data["class_names"][0]
         return {
-            # Format class names appropriately for directory creation.
             class_name[0].replace(" ", "_").replace("/", "_"): i
             for i, class_name in enumerate(class_names)
         }
 
-    def _get_annotations(self):
+    def _make_dataset(self):
         """
         Returns Annotations for training data and testing data
         """
         annotations = None
         if self.train:
-            annotations = io.loadmat(os.path.join(self.root, "devkit/cars_train_annos.mat"))
+            annotations = self._loadmat(os.path.join(self.root, "devkit/cars_train_annos.mat"))
         else:
-            annotations = io.loadmat(os.path.join(self.root, "cars_test_annos_withlabels.mat"))
-        return annotations["annotations"][0]
+            annotations = self._loadmat(os.path.join(self.root, "cars_test_annos_withlabels.mat"))
+        samples = []
+        annotations = annotations["annotations"][0]
+        for index in range(len(annotations)):
+            target = annotations[index][4][0, 0]
+            image_file = annotations[index][5][0]
+            samples.append((image_file, target))
+        return samples
 
     def __len__(self) -> int:
-        return self.annotations.shape[0]
+        return len(self._samples)
 
-    def __getitem__(self, idx: int) -> (Image,int):
+    def __getitem__(self, idx: int) -> (Image, int):
         """Returns pil_image and class_id for given index"""
-        target_id = self.annotations[idx][4][0, 0]
-        image_name = self.annotations[idx][5][0]
-        # 0,1,2,3 correspond to bounding boxes
-
-        image_path = os.path.join(self.root, f"cars_{'train' if self.train else 'test' }", image_name)
+        image_file, target = self._samples[idx]
+        image_path = os.path.join(self.root, f"cars_{'train' if self.train else 'test'}", image_file)
         # Beware: Stanford cars targets starts at 1
-        target_id = target_id - 1
+        target_id = target - 1
         pil_image = Image.open(image_path).convert("RGB")
 
         if self.transform is not None:
@@ -106,7 +115,6 @@ def __getitem__(self, idx: int) -> (Image,int):
             pil_image = self.target_transform(pil_image)
         return pil_image, target_id
 
-
     def download(self) -> None:
         if self._check_exists():
             return
@@ -118,11 +126,16 @@ def download(self) -> None:
                 md5=self.md5s[self.train]
             )
             download_and_extract_archive(
-                    url=self.annot_urls[1], download_root=self.root, extract_root=self.root ,md5="c3b158d763b6e2245038c8ad08e45376"
-                )
+                url=self.annot_urls[1], download_root=self.root, extract_root=self.root,
+                md5=self.annot_md5s[1]
+            )
             if not self.train:
                 download_url(
-                    url=self.annot_urls[0], filename="cars_test_annos_withlabels.mat", root=self.root ,md5="b0a2b23655a3edd16d84508592a98d10"
+                    url=self.annot_urls[0], filename="cars_test_annos_withlabels.mat", root=self.root,
+                    md5=self.annot_md5s[0]
                 )
+
     def _check_exists(self) -> bool:
-        return os.path.exists(os.path.join(self.root, f"cars_{'train' if self.train else 'test'}")) and os.path.isdir(os.path.join(self.root, f"cars_{'train' if self.train else 'test'}")) and os.path.exists(os.path.join(self.root,"devkit/cars_meta.mat"))
\ No newline at end of file
+        return os.path.exists(os.path.join(self.root, f"cars_{'train' if self.train else 'test'}")) and os.path.isdir(
+            os.path.join(self.root, f"cars_{'train' if self.train else 'test'}")) and os.path.exists(
+            os.path.join(self.root, "devkit/cars_meta.mat")) if self.train else os.path.exists(os.path.join(self.root,"cars_test_annos_withlabels.mat"))

From db410b91a43f65c0b38cd749e1e6bbd33a6ad0ce Mon Sep 17 00:00:00 2001
From: Abhijit Deo <72816663+abhi-glitchhg@users.noreply.github.com>
Date: Sat, 8 Jan 2022 14:25:58 +0530
Subject: [PATCH 06/15] Adding Testcase for stanford cars

---
 test/test_datasets.py                  | 41 ++++++++++++++++++++++++++
 torchvision/datasets/standford_cars.py | 10 +++----
 2 files changed, 46 insertions(+), 5 deletions(-)

diff --git a/test/test_datasets.py b/test/test_datasets.py
index 6ba39542236..3add5516d24 100644
--- a/test/test_datasets.py
+++ b/test/test_datasets.py
@@ -2326,6 +2326,35 @@ class StanfordCarsTestCase(datasets_utils.ImageDatasetTestCase):
     FEATURE_TYPES = (PIL.Image.Image, int)
     ADDITIONAL_CONFIGS = datasets_utils.combinations_grid(train=(True,False))
 
+    def _inject_fake_data(self, tmpdir, config):
+        root_folder = os.path.join(tmpdir, "stanforddcars")
+
+        os.makedirs(root_folder, exist_ok=True)
+
+        num_examples = 5
+        if config["train"]:
+            #if train = True
+            datasets_utils.create_image_folder(
+            root=root_folder,
+            name="cars_train",
+            file_name_fn=lambda image_index: f"{image_index:5d}.jpg",
+            num_examples=num_examples,
+            )
+
+
+        else:
+            # test_folder i.e train = False
+            datasets_utils.create_image_folder(
+            root_folder,
+            "cars_test",
+            lambda image_index: f"{image_index:5d}.jpg",
+            num_examples
+            )
+
+            file = "cars_test_annos_withlabels.mat"
+
+
+
 
 
 class GTSRBTestCase(datasets_utils.ImageDatasetTestCase):
@@ -2469,6 +2498,18 @@ def _meta_to_split_and_classification_ann(self, meta, idx):
         breed_id = "-1"
         return (image_id, class_id, species, breed_id)
 
+class StanfordCarsTestCase(datasets_utils.ImageDatasetTestCase):
+    DATASET_CLASS = datasets.StanfordCars
+    REQUIRED_PACKAGES = ("scipy",)
+    ADDITIONAL_CONFIGS = datasets_utils.combinations_grid(train=(True,False))
+    FEATURE_TYPES = (PIL.Image.Image,int)
+
+    def _inject_fake_data(self, tmpdir, config):
+
+        import scipy.io as io
+        train = config["train"]
+        num_examples= 5
+        pass
 
 if __name__ == "__main__":
     unittest.main()
diff --git a/torchvision/datasets/standford_cars.py b/torchvision/datasets/standford_cars.py
index 3d13e5b07b3..94daeac8279 100644
--- a/torchvision/datasets/standford_cars.py
+++ b/torchvision/datasets/standford_cars.py
@@ -68,7 +68,7 @@ def __init__(
             raise RuntimeError("Dataset not found. You can use download=True to download it")
 
         self._samples = self._make_dataset()
-        self.classes = self._get_class_names()  # class_id to class_name mapping
+        self.classes = self._get_classes_name()  # class_id to class_name mapping
 
     def _get_class_names(self) -> dict:
         """
@@ -95,6 +95,8 @@ def _make_dataset(self):
         for index in range(len(annotations)):
             target = annotations[index][4][0, 0]
             image_file = annotations[index][5][0]
+            # Beware: Stanford cars targets starts at 1
+            target = target - 1
             samples.append((image_file, target))
         return samples
 
@@ -105,15 +107,13 @@ def __getitem__(self, idx: int) -> (Image, int):
         """Returns pil_image and class_id for given index"""
         image_file, target = self._samples[idx]
         image_path = os.path.join(self.root, f"cars_{'train' if self.train else 'test'}", image_file)
-        # Beware: Stanford cars targets starts at 1
-        target_id = target - 1
         pil_image = Image.open(image_path).convert("RGB")
 
         if self.transform is not None:
             pil_image = self.transform(pil_image)
         if self.target_transform is not None:
-            pil_image = self.target_transform(pil_image)
-        return pil_image, target_id
+            target = self.target_transform(target)
+        return pil_image, target
 
     def download(self) -> None:
         if self._check_exists():

From 0e91bf0454f35a35a58e0826d50a49e837a79d96 Mon Sep 17 00:00:00 2001
From: Abhijit Deo <72816663+abhi-glitchhg@users.noreply.github.com>
Date: Sat, 8 Jan 2022 15:30:44 +0530
Subject: [PATCH 07/15] Added Testcase for stanford cars

---
 test/test_datasets.py                  | 42 ++++++++++++++++++-
 torchvision/datasets/standford_cars.py | 56 ++++++++++++++------------
 2 files changed, 70 insertions(+), 28 deletions(-)

diff --git a/test/test_datasets.py b/test/test_datasets.py
index 3add5516d24..98cdfadaee4 100644
--- a/test/test_datasets.py
+++ b/test/test_datasets.py
@@ -2498,6 +2498,8 @@ def _meta_to_split_and_classification_ann(self, meta, idx):
         breed_id = "-1"
         return (image_id, class_id, species, breed_id)
 
+
+
 class StanfordCarsTestCase(datasets_utils.ImageDatasetTestCase):
     DATASET_CLASS = datasets.StanfordCars
     REQUIRED_PACKAGES = ("scipy",)
@@ -2505,11 +2507,47 @@ class StanfordCarsTestCase(datasets_utils.ImageDatasetTestCase):
     FEATURE_TYPES = (PIL.Image.Image,int)
 
     def _inject_fake_data(self, tmpdir, config):
-
         import scipy.io as io
+        from numpy.core.records import fromarrays #to record arrays similar to matlab format
         train = config["train"]
         num_examples= 5
-        pass
+        root_folder = os.path.join(tmpdir, "stanforddcars")
+        os.makedirs(root_folder, exist_ok=True)
+
+        #generate random data for labels
+        class_name = np.random.randint(0, 100, num_examples, dtype=np.uint8)
+        bbox_x1 = np.random.randint(0, 100, num_examples, dtype=np.uint8)
+        bbox_x2 = np.random.randint(0, 100, num_examples, dtype=np.uint8)
+
+        bbox_y1 = np.random.randint(0, 100, num_examples, dtype=np.uint8)
+        bb1ox_y2= np.random.randint(0, 100, num_examples, dtype=np.uint8)
+        fname = [f"{i:5d}.jpg" for i in range(num_examples)]
+
+        rec_array = fromarrays([bbox_x1,bbox_y1,bbox_x2,bb1ox_y2,class_name,fname], ["bbox_x1","bbox_y1","bbox_x2","bbox_y2","class","fname"])
+
+        if train:
+            #create training image folder
+            datasets_utils.create_image_folder(
+            root=root_folder,
+            name="cars_train",
+            file_name_fn=lambda  image_index : f"{image_index:5d}.jpg",
+            num_examples=num_examples,
+            )
+            devkit = os.path.join(root_folder,"devkit")
+            io.savemat(f'{devkit}/cars_train_annos.mat', {'annotations': rec_array}) #save the recorded array as matlab file
+        else:
+            #create test image folder
+            datasets_utils.create_image_folder(
+            root=root_folder,
+            name="cars_test",
+            file_name_fn=lambda  image_index : f"{image_index:5d}.jpg",
+            num_examples=num_examples
+            )
+            io.savemat(f"{root_folder}/cars_test_annos_withlabels.mat", {"annotations":rec_array}) #save recorded array as matlab file
+
+        return num_examples
+
+
 
 if __name__ == "__main__":
     unittest.main()
diff --git a/torchvision/datasets/standford_cars.py b/torchvision/datasets/standford_cars.py
index 94daeac8279..b319ee702f2 100644
--- a/torchvision/datasets/standford_cars.py
+++ b/torchvision/datasets/standford_cars.py
@@ -1,6 +1,7 @@
 import os
 import os.path
 from typing import Callable, Optional
+
 from PIL import Image
 
 from .utils import download_and_extract_archive, download_url
@@ -16,7 +17,7 @@ class StanfordCars(VisionDataset):
 
     Args:
         root (string): Root directory of dataset
-        train (bool, optional):
+        train (bool, optional):If True, creates dataset from training set, otherwise creates from test set
         transform (callable, optional): A function/transform that  takes in an PIL image
             and returns a transformed version. E.g, ``transforms.RandomCrop``
         target_transform (callable, optional): A function/transform that takes in the
@@ -28,26 +29,30 @@ class StanfordCars(VisionDataset):
     urls = (
         "https://ai.stanford.edu/~jkrause/car196/cars_test.tgz",
         "https://ai.stanford.edu/~jkrause/car196/cars_train.tgz",
-    )
+    )  # test and train image urls
 
-    md5s = ("4ce7ebf6a94d07f1952d94dd34c4d501", "065e5b463ae28d29e77c1b4b166cfe61")  # md5checksum for test and train
+    md5s = (
+        "4ce7ebf6a94d07f1952d94dd34c4d501",
+        "065e5b463ae28d29e77c1b4b166cfe61",
+    )  # md5checksum for test and train data
 
     annot_urls = (
         "https://ai.stanford.edu/~jkrause/car196/cars_test_annos_withlabels.mat",
         "https://ai.stanford.edu/~jkrause/cars/car_devkit.tgz",
-    )
+    )  # annotations and labels for test and train
+
     annot_md5s = (
         "b0a2b23655a3edd16d84508592a98d10",
         "c3b158d763b6e2245038c8ad08e45376",
-    )
+    )  # md5 checksum for annotations
 
     def __init__(
-            self,
-            root: str,
-            train: bool = True,
-            transform: Optional[Callable] = None,
-            target_transform: Optional[Callable] = None,
-            download: bool = False,
+        self,
+        root: str,
+        train: bool = True,
+        transform: Optional[Callable] = None,
+        target_transform: Optional[Callable] = None,
+        download: bool = False,
     ) -> None:
 
         try:
@@ -76,10 +81,7 @@ def _get_class_names(self) -> dict:
         """
         meta_data = self._loadmat(os.path.join(self.root, "devkit/cars_meta.mat"))
         class_names = meta_data["class_names"][0]
-        return {
-            class_name[0].replace(" ", "_").replace("/", "_"): i
-            for i, class_name in enumerate(class_names)
-        }
+        return {class_name[0].replace(" ", "_").replace("/", "_"): i for i, class_name in enumerate(class_names)}
 
     def _make_dataset(self):
         """
@@ -120,22 +122,24 @@ def download(self) -> None:
             return
         else:
             download_and_extract_archive(
-                url=self.urls[self.train],
-                download_root=self.root,
-                extract_root=self.root,
-                md5=self.md5s[self.train]
+                url=self.urls[self.train], download_root=self.root, extract_root=self.root, md5=self.md5s[self.train]
             )
             download_and_extract_archive(
-                url=self.annot_urls[1], download_root=self.root, extract_root=self.root,
-                md5=self.annot_md5s[1]
+                url=self.annot_urls[1], download_root=self.root, extract_root=self.root, md5=self.annot_md5s[1]
             )
             if not self.train:
                 download_url(
-                    url=self.annot_urls[0], filename="cars_test_annos_withlabels.mat", root=self.root,
-                    md5=self.annot_md5s[0]
+                    url=self.annot_urls[0],
+                    filename="cars_test_annos_withlabels.mat",
+                    root=self.root,
+                    md5=self.annot_md5s[0],
                 )
 
     def _check_exists(self) -> bool:
-        return os.path.exists(os.path.join(self.root, f"cars_{'train' if self.train else 'test'}")) and os.path.isdir(
-            os.path.join(self.root, f"cars_{'train' if self.train else 'test'}")) and os.path.exists(
-            os.path.join(self.root, "devkit/cars_meta.mat")) if self.train else os.path.exists(os.path.join(self.root,"cars_test_annos_withlabels.mat"))
+        return (
+            os.path.exists(os.path.join(self.root, f"cars_{'train' if self.train else 'test'}"))
+            and os.path.isdir(os.path.join(self.root, f"cars_{'train' if self.train else 'test'}"))
+            and os.path.exists(os.path.join(self.root, "devkit/cars_meta.mat"))
+            if self.train
+            else os.path.exists(os.path.join(self.root, "cars_test_annos_withlabels.mat"))
+        )

From fbd3122c48d8e3ceb09f9a853cb0e56b2feb068c Mon Sep 17 00:00:00 2001
From: Abhijit Deo <72816663+abhi-glitchhg@users.noreply.github.com>
Date: Sat, 8 Jan 2022 16:11:54 +0530
Subject: [PATCH 08/15] Added Testcase for stanford cars

---
 test/test_datasets.py | 86 ++++++++++++++-----------------------------
 1 file changed, 28 insertions(+), 58 deletions(-)

diff --git a/test/test_datasets.py b/test/test_datasets.py
index 98cdfadaee4..5a1085dfa01 100644
--- a/test/test_datasets.py
+++ b/test/test_datasets.py
@@ -2320,42 +2320,6 @@ def inject_fake_data(self, tmpdir, config):
 
         return num_samples
 
-class StanfordCarsTestCase(datasets_utils.ImageDatasetTestCase):
-    DATASET_CLASS = datasets.StanfordCars
-    REQUIRED_PACKAGES = ("scipy",)
-    FEATURE_TYPES = (PIL.Image.Image, int)
-    ADDITIONAL_CONFIGS = datasets_utils.combinations_grid(train=(True,False))
-
-    def _inject_fake_data(self, tmpdir, config):
-        root_folder = os.path.join(tmpdir, "stanforddcars")
-
-        os.makedirs(root_folder, exist_ok=True)
-
-        num_examples = 5
-        if config["train"]:
-            #if train = True
-            datasets_utils.create_image_folder(
-            root=root_folder,
-            name="cars_train",
-            file_name_fn=lambda image_index: f"{image_index:5d}.jpg",
-            num_examples=num_examples,
-            )
-
-
-        else:
-            # test_folder i.e train = False
-            datasets_utils.create_image_folder(
-            root_folder,
-            "cars_test",
-            lambda image_index: f"{image_index:5d}.jpg",
-            num_examples
-            )
-
-            file = "cars_test_annos_withlabels.mat"
-
-
-
-
 
 class GTSRBTestCase(datasets_utils.ImageDatasetTestCase):
     DATASET_CLASS = datasets.GTSRB
@@ -2499,55 +2463,61 @@ def _meta_to_split_and_classification_ann(self, meta, idx):
         return (image_id, class_id, species, breed_id)
 
 
-
 class StanfordCarsTestCase(datasets_utils.ImageDatasetTestCase):
     DATASET_CLASS = datasets.StanfordCars
     REQUIRED_PACKAGES = ("scipy",)
-    ADDITIONAL_CONFIGS = datasets_utils.combinations_grid(train=(True,False))
-    FEATURE_TYPES = (PIL.Image.Image,int)
+    ADDITIONAL_CONFIGS = datasets_utils.combinations_grid(train=(True, False))
+    FEATURE_TYPES = (PIL.Image.Image, int)
 
     def _inject_fake_data(self, tmpdir, config):
         import scipy.io as io
-        from numpy.core.records import fromarrays #to record arrays similar to matlab format
+        from numpy.core.records import fromarrays  # to record arrays similar to matlab format
+
         train = config["train"]
-        num_examples= 5
+        num_examples = 5
         root_folder = os.path.join(tmpdir, "stanforddcars")
         os.makedirs(root_folder, exist_ok=True)
 
-        #generate random data for labels
+        # generate random data for labels
         class_name = np.random.randint(0, 100, num_examples, dtype=np.uint8)
         bbox_x1 = np.random.randint(0, 100, num_examples, dtype=np.uint8)
         bbox_x2 = np.random.randint(0, 100, num_examples, dtype=np.uint8)
 
         bbox_y1 = np.random.randint(0, 100, num_examples, dtype=np.uint8)
-        bb1ox_y2= np.random.randint(0, 100, num_examples, dtype=np.uint8)
+        bb1ox_y2 = np.random.randint(0, 100, num_examples, dtype=np.uint8)
         fname = [f"{i:5d}.jpg" for i in range(num_examples)]
 
-        rec_array = fromarrays([bbox_x1,bbox_y1,bbox_x2,bb1ox_y2,class_name,fname], ["bbox_x1","bbox_y1","bbox_x2","bbox_y2","class","fname"])
+        rec_array = fromarrays(
+            [bbox_x1, bbox_y1, bbox_x2, bb1ox_y2, class_name, fname],
+            ["bbox_x1", "bbox_y1", "bbox_x2", "bbox_y2", "class", "fname"],
+        )
 
         if train:
-            #create training image folder
+            # create training image folder
             datasets_utils.create_image_folder(
-            root=root_folder,
-            name="cars_train",
-            file_name_fn=lambda  image_index : f"{image_index:5d}.jpg",
-            num_examples=num_examples,
+                root=root_folder,
+                name="cars_train",
+                file_name_fn=lambda image_index: f"{image_index:5d}.jpg",
+                num_examples=num_examples,
             )
-            devkit = os.path.join(root_folder,"devkit")
-            io.savemat(f'{devkit}/cars_train_annos.mat', {'annotations': rec_array}) #save the recorded array as matlab file
+            devkit = os.path.join(root_folder, "devkit")
+            io.savemat(
+                f"{devkit}/cars_train_annos.mat", {"annotations": rec_array}
+            )  # save the recorded array as matlab file
         else:
-            #create test image folder
+            # create test image folder
             datasets_utils.create_image_folder(
-            root=root_folder,
-            name="cars_test",
-            file_name_fn=lambda  image_index : f"{image_index:5d}.jpg",
-            num_examples=num_examples
+                root=root_folder,
+                name="cars_test",
+                file_name_fn=lambda image_index: f"{image_index:5d}.jpg",
+                num_examples=num_examples,
             )
-            io.savemat(f"{root_folder}/cars_test_annos_withlabels.mat", {"annotations":rec_array}) #save recorded array as matlab file
+            io.savemat(
+                f"{root_folder}/cars_test_annos_withlabels.mat", {"annotations": rec_array}
+            )  # save recorded array as matlab file
 
         return num_examples
 
 
-
 if __name__ == "__main__":
     unittest.main()

From 817e9f240c5bb6908bae059574f3b5347609fd2e Mon Sep 17 00:00:00 2001
From: Abhijit Deo <72816663+abhi-glitchhg@users.noreply.github.com>
Date: Sat, 8 Jan 2022 22:24:22 +0530
Subject: [PATCH 09/15] minor edit

---
 test/test_datasets.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/test/test_datasets.py b/test/test_datasets.py
index 5a1085dfa01..55eb0794e09 100644
--- a/test/test_datasets.py
+++ b/test/test_datasets.py
@@ -2489,7 +2489,7 @@ def _inject_fake_data(self, tmpdir, config):
 
         rec_array = fromarrays(
             [bbox_x1, bbox_y1, bbox_x2, bb1ox_y2, class_name, fname],
-            ["bbox_x1", "bbox_y1", "bbox_x2", "bbox_y2", "class", "fname"],
+            names=["bbox_x1", "bbox_y1", "bbox_x2", "bbox_y2", "class", "fname"],
         )
 
         if train:

From 52c98f34fdb6b2c986893c04cefbe9b4b1f4be4d Mon Sep 17 00:00:00 2001
From: Abhijit Deo <72816663+abhi-glitchhg@users.noreply.github.com>
Date: Wed, 12 Jan 2022 00:09:37 +0530
Subject: [PATCH 10/15] made changes as per the suggestions

---
 test/test_datasets.py                  |  4 +-
 torchvision/datasets/standford_cars.py | 79 ++++++++++----------------
 2 files changed, 31 insertions(+), 52 deletions(-)

diff --git a/test/test_datasets.py b/test/test_datasets.py
index 55eb0794e09..44d40c6e098 100644
--- a/test/test_datasets.py
+++ b/test/test_datasets.py
@@ -2467,7 +2467,6 @@ class StanfordCarsTestCase(datasets_utils.ImageDatasetTestCase):
     DATASET_CLASS = datasets.StanfordCars
     REQUIRED_PACKAGES = ("scipy",)
     ADDITIONAL_CONFIGS = datasets_utils.combinations_grid(train=(True, False))
-    FEATURE_TYPES = (PIL.Image.Image, int)
 
     def _inject_fake_data(self, tmpdir, config):
         import scipy.io as io
@@ -2475,7 +2474,7 @@ def _inject_fake_data(self, tmpdir, config):
 
         train = config["train"]
         num_examples = 5
-        root_folder = os.path.join(tmpdir, "stanforddcars")
+        root_folder = os.path.join(tmpdir)
         os.makedirs(root_folder, exist_ok=True)
 
         # generate random data for labels
@@ -2501,6 +2500,7 @@ def _inject_fake_data(self, tmpdir, config):
                 num_examples=num_examples,
             )
             devkit = os.path.join(root_folder, "devkit")
+            os.makedirs(devkit)
             io.savemat(
                 f"{devkit}/cars_train_annos.mat", {"annotations": rec_array}
             )  # save the recorded array as matlab file
diff --git a/torchvision/datasets/standford_cars.py b/torchvision/datasets/standford_cars.py
index b319ee702f2..8dcd8b7c6a6 100644
--- a/torchvision/datasets/standford_cars.py
+++ b/torchvision/datasets/standford_cars.py
@@ -1,6 +1,6 @@
 import os
 import os.path
-from typing import Callable, Optional
+from typing import Callable, Optional, Any, Tuple
 
 from PIL import Image
 
@@ -56,9 +56,7 @@ def __init__(
     ) -> None:
 
         try:
-            from scipy.io import loadmat
-
-            self._loadmat = loadmat
+            import scipy.io as sio
         except ImportError:
             raise RuntimeError("Scipy is not found. This dataset needs to have scipy installed: pip install scipy")
 
@@ -72,43 +70,29 @@ def __init__(
         if not self._check_exists():
             raise RuntimeError("Dataset not found. You can use download=True to download it")
 
-        self._samples = self._make_dataset()
-        self.classes = self._get_classes_name()  # class_id to class_name mapping
-
-    def _get_class_names(self) -> dict:
-        """
-        Returns Mapping of class ids to class names in form of Dictionary
-        """
-        meta_data = self._loadmat(os.path.join(self.root, "devkit/cars_meta.mat"))
-        class_names = meta_data["class_names"][0]
-        return {class_name[0].replace(" ", "_").replace("/", "_"): i for i, class_name in enumerate(class_names)}
-
-    def _make_dataset(self):
-        """
-        Returns Annotations for training data and testing data
-        """
-        annotations = None
-        if self.train:
-            annotations = self._loadmat(os.path.join(self.root, "devkit/cars_train_annos.mat"))
-        else:
-            annotations = self._loadmat(os.path.join(self.root, "cars_test_annos_withlabels.mat"))
-        samples = []
-        annotations = annotations["annotations"][0]
-        for index in range(len(annotations)):
-            target = annotations[index][4][0, 0]
-            image_file = annotations[index][5][0]
-            # Beware: Stanford cars targets starts at 1
-            target = target - 1
-            samples.append((image_file, target))
-        return samples
+        self._samples = [
+            (
+                os.path.join(self.root, f"cars_{'train' if self.train else 'test'}", annotation["fname"]),
+                annotation["class"] - 1,  # Beware stanford cars target mapping  starts from 1
+            )
+            for annotation in sio.loadmat(
+                os.path.join(
+                    self.root,
+                    *["devkit", "cars_train_annos.mat"] if self.train else ["cars_test_annos_withlabels.mat"],
+                ),
+                squeeze_me=True,
+            )["annotations"]
+        ]
+
+        class_names = sio.loadmat(os.path.join(self.root, "devkit", "cars_meta.mat"))["class_names"][0]
+        self.classes = {class_name[0]: i for i, class_name in enumerate(class_names)}
 
     def __len__(self) -> int:
         return len(self._samples)
 
-    def __getitem__(self, idx: int) -> (Image, int):
+    def __getitem__(self, idx: int) -> Tuple[Any, Any]:
         """Returns pil_image and class_id for given index"""
-        image_file, target = self._samples[idx]
-        image_path = os.path.join(self.root, f"cars_{'train' if self.train else 'test'}", image_file)
+        image_path, target = self._samples[idx]
         pil_image = Image.open(image_path).convert("RGB")
 
         if self.transform is not None:
@@ -120,26 +104,21 @@ def __getitem__(self, idx: int) -> (Image, int):
     def download(self) -> None:
         if self._check_exists():
             return
-        else:
-            download_and_extract_archive(
-                url=self.urls[self.train], download_root=self.root, extract_root=self.root, md5=self.md5s[self.train]
-            )
-            download_and_extract_archive(
-                url=self.annot_urls[1], download_root=self.root, extract_root=self.root, md5=self.annot_md5s[1]
+
+        download_and_extract_archive(url=self.urls[self.train], download_root=self.root, md5=self.md5s[self.train])
+        download_and_extract_archive(url=self.annot_urls[1], download_root=self.root, md5=self.annot_md5s[1])
+        if not self.train:
+            download_url(
+                url=self.annot_urls[0],
+                root=self.root,
+                md5=self.annot_md5s[0],
             )
-            if not self.train:
-                download_url(
-                    url=self.annot_urls[0],
-                    filename="cars_test_annos_withlabels.mat",
-                    root=self.root,
-                    md5=self.annot_md5s[0],
-                )
 
     def _check_exists(self) -> bool:
         return (
             os.path.exists(os.path.join(self.root, f"cars_{'train' if self.train else 'test'}"))
             and os.path.isdir(os.path.join(self.root, f"cars_{'train' if self.train else 'test'}"))
-            and os.path.exists(os.path.join(self.root, "devkit/cars_meta.mat"))
+            and os.path.exists(os.path.join(self.root, "devkit", "cars_meta.mat"))
             if self.train
             else os.path.exists(os.path.join(self.root, "cars_test_annos_withlabels.mat"))
         )

From 2eceade03a79f987593475e618f504c7afe5f0ba Mon Sep 17 00:00:00 2001
From: Abhijit Deo <72816663+abhi-glitchhg@users.noreply.github.com>
Date: Wed, 12 Jan 2022 00:41:20 +0530
Subject: [PATCH 11/15] fixed typo in naming stanford_cars.py

---
 torchvision/datasets/__init__.py                             | 2 +-
 torchvision/datasets/{standford_cars.py => stanford_cars.py} | 0
 2 files changed, 1 insertion(+), 1 deletion(-)
 rename torchvision/datasets/{standford_cars.py => stanford_cars.py} (100%)

diff --git a/torchvision/datasets/__init__.py b/torchvision/datasets/__init__.py
index 9a355f7ba32..480ee22d58a 100644
--- a/torchvision/datasets/__init__.py
+++ b/torchvision/datasets/__init__.py
@@ -27,7 +27,7 @@
 from .sbd import SBDataset
 from .sbu import SBU
 from .semeion import SEMEION
-from .standford_cars import StanfordCars
+from .stanford_cars import StanfordCars
 from .stl10 import STL10
 from .sun397 import SUN397
 from .svhn import SVHN
diff --git a/torchvision/datasets/standford_cars.py b/torchvision/datasets/stanford_cars.py
similarity index 100%
rename from torchvision/datasets/standford_cars.py
rename to torchvision/datasets/stanford_cars.py

From af25d727071d0c2eb870d4013c1fccd3521e706e Mon Sep 17 00:00:00 2001
From: Abhijit Deo <72816663+abhi-glitchhg@users.noreply.github.com>
Date: Mon, 17 Jan 2022 14:07:50 +0530
Subject: [PATCH 12/15] cars_meta.mat file will be created in test

---
 test/test_datasets.py                 | 26 ++++++++++++--------------
 torchvision/datasets/stanford_cars.py |  9 ++++++---
 2 files changed, 18 insertions(+), 17 deletions(-)

diff --git a/test/test_datasets.py b/test/test_datasets.py
index 44d40c6e098..5f3dd42a90f 100644
--- a/test/test_datasets.py
+++ b/test/test_datasets.py
@@ -2470,14 +2470,12 @@ class StanfordCarsTestCase(datasets_utils.ImageDatasetTestCase):
 
     def _inject_fake_data(self, tmpdir, config):
         import scipy.io as io
-        from numpy.core.records import fromarrays  # to record arrays similar to matlab format
+        from numpy.core.records import fromarrays
 
         train = config["train"]
         num_examples = 5
-        root_folder = os.path.join(tmpdir)
-        os.makedirs(root_folder, exist_ok=True)
+        root_folder = tmpdir
 
-        # generate random data for labels
         class_name = np.random.randint(0, 100, num_examples, dtype=np.uint8)
         bbox_x1 = np.random.randint(0, 100, num_examples, dtype=np.uint8)
         bbox_x2 = np.random.randint(0, 100, num_examples, dtype=np.uint8)
@@ -2490,31 +2488,31 @@ def _inject_fake_data(self, tmpdir, config):
             [bbox_x1, bbox_y1, bbox_x2, bb1ox_y2, class_name, fname],
             names=["bbox_x1", "bbox_y1", "bbox_x2", "bbox_y2", "class", "fname"],
         )
+        devkit = os.path.join(root_folder, "devkit")
+        os.makedirs(devkit)
+
+        random_class_names = ["Tesla Model S Sedan 2012"] * 196
+
+        io.savemat(os.path.join(devkit, "cars_meta.mat"), {"class_names": random_class_names})
 
         if train:
-            # create training image folder
             datasets_utils.create_image_folder(
                 root=root_folder,
                 name="cars_train",
                 file_name_fn=lambda image_index: f"{image_index:5d}.jpg",
                 num_examples=num_examples,
             )
-            devkit = os.path.join(root_folder, "devkit")
-            os.makedirs(devkit)
-            io.savemat(
-                f"{devkit}/cars_train_annos.mat", {"annotations": rec_array}
-            )  # save the recorded array as matlab file
+
+            io.savemat(f"{devkit}/cars_train_annos.mat", {"annotations": rec_array})
         else:
-            # create test image folder
+
             datasets_utils.create_image_folder(
                 root=root_folder,
                 name="cars_test",
                 file_name_fn=lambda image_index: f"{image_index:5d}.jpg",
                 num_examples=num_examples,
             )
-            io.savemat(
-                f"{root_folder}/cars_test_annos_withlabels.mat", {"annotations": rec_array}
-            )  # save recorded array as matlab file
+            io.savemat(f"{root_folder}/cars_test_annos_withlabels.mat", {"annotations": rec_array})
 
         return num_examples
 
diff --git a/torchvision/datasets/stanford_cars.py b/torchvision/datasets/stanford_cars.py
index 8dcd8b7c6a6..9af9dc3fb49 100644
--- a/torchvision/datasets/stanford_cars.py
+++ b/torchvision/datasets/stanford_cars.py
@@ -73,7 +73,8 @@ def __init__(
         self._samples = [
             (
                 os.path.join(self.root, f"cars_{'train' if self.train else 'test'}", annotation["fname"]),
-                annotation["class"] - 1,  # Beware stanford cars target mapping  starts from 1
+                annotation["class"] - 1,
+                # Beware stanford cars target mapping  starts from 1
             )
             for annotation in sio.loadmat(
                 os.path.join(
@@ -84,8 +85,10 @@ def __init__(
             )["annotations"]
         ]
 
-        class_names = sio.loadmat(os.path.join(self.root, "devkit", "cars_meta.mat"))["class_names"][0]
-        self.classes = {class_name[0]: i for i, class_name in enumerate(class_names)}
+        self.classes = sio.loadmat(os.path.join(self.root, "devkit", "cars_meta.mat"), squeeze_me=True)[
+            "class_names"
+        ].tolist()
+        self.class_to_idx = {cls: i for i, cls in enumerate(self.classes)}
 
     def __len__(self) -> int:
         return len(self._samples)

From 8fceb0b661f2988ee2c39f30834368d1b72fd98b Mon Sep 17 00:00:00 2001
From: Nicolas Hug <nicolashug@fb.com>
Date: Mon, 17 Jan 2022 17:57:56 +0000
Subject: [PATCH 13/15] Some cleanups

---
 test/test_datasets.py                 |  64 +++++++----------
 torchvision/datasets/stanford_cars.py | 100 ++++++++++++--------------
 2 files changed, 74 insertions(+), 90 deletions(-)

diff --git a/test/test_datasets.py b/test/test_datasets.py
index dad536d48e1..8a5dba326ff 100644
--- a/test/test_datasets.py
+++ b/test/test_datasets.py
@@ -2517,53 +2517,43 @@ def _meta_to_split_and_classification_ann(self, meta, idx):
 class StanfordCarsTestCase(datasets_utils.ImageDatasetTestCase):
     DATASET_CLASS = datasets.StanfordCars
     REQUIRED_PACKAGES = ("scipy",)
-    ADDITIONAL_CONFIGS = datasets_utils.combinations_grid(train=(True, False))
+    ADDITIONAL_CONFIGS = datasets_utils.combinations_grid(split=("train", "test"))
 
-    def _inject_fake_data(self, tmpdir, config):
+    def inject_fake_data(self, tmpdir, config):
         import scipy.io as io
         from numpy.core.records import fromarrays
 
-        train = config["train"]
-        num_examples = 5
-        root_folder = tmpdir
+        num_examples = {"train": 5, "test": 7}[config["split"]]
+        num_classes = 3
+        base_folder = pathlib.Path(tmpdir) / "stanford_cars"
 
-        class_name = np.random.randint(0, 100, num_examples, dtype=np.uint8)
-        bbox_x1 = np.random.randint(0, 100, num_examples, dtype=np.uint8)
-        bbox_x2 = np.random.randint(0, 100, num_examples, dtype=np.uint8)
+        devkit = base_folder / "devkit"
+        devkit.mkdir(parents=True)
 
-        bbox_y1 = np.random.randint(0, 100, num_examples, dtype=np.uint8)
-        bb1ox_y2 = np.random.randint(0, 100, num_examples, dtype=np.uint8)
-        fname = [f"{i:5d}.jpg" for i in range(num_examples)]
+        if config["split"] == "train":
+            images_folder_name = "cars_train"
+            annotations_mat_path = str(devkit / "cars_train_annos.mat")
+        else:
+            images_folder_name = "cars_test"
+            annotations_mat_path = str(base_folder / "cars_test_annos_withlabels.mat")
 
-        rec_array = fromarrays(
-            [bbox_x1, bbox_y1, bbox_x2, bb1ox_y2, class_name, fname],
-            names=["bbox_x1", "bbox_y1", "bbox_x2", "bbox_y2", "class", "fname"],
+        datasets_utils.create_image_folder(
+            root=base_folder,
+            name=images_folder_name,
+            file_name_fn=lambda image_index: f"{image_index:5d}.jpg",
+            num_examples=num_examples,
         )
-        devkit = os.path.join(root_folder, "devkit")
-        os.makedirs(devkit)
-
-        random_class_names = ["Tesla Model S Sedan 2012"] * 196
-
-        io.savemat(os.path.join(devkit, "cars_meta.mat"), {"class_names": random_class_names})
 
-        if train:
-            datasets_utils.create_image_folder(
-                root=root_folder,
-                name="cars_train",
-                file_name_fn=lambda image_index: f"{image_index:5d}.jpg",
-                num_examples=num_examples,
-            )
-
-            io.savemat(f"{devkit}/cars_train_annos.mat", {"annotations": rec_array})
-        else:
+        classes = np.random.randint(1, num_classes + 1, num_examples, dtype=np.uint8)
+        fnames = [f"{i:5d}.jpg" for i in range(num_examples)]
+        rec_array = fromarrays(
+            [classes, fnames],
+            names=["class", "fname"],
+        )
+        io.savemat(annotations_mat_path, {"annotations": rec_array})
 
-            datasets_utils.create_image_folder(
-                root=root_folder,
-                name="cars_test",
-                file_name_fn=lambda image_index: f"{image_index:5d}.jpg",
-                num_examples=num_examples,
-            )
-            io.savemat(f"{root_folder}/cars_test_annos_withlabels.mat", {"annotations": rec_array})
+        random_class_names = ["random_name"] * num_classes
+        io.savemat(str(devkit / "cars_meta.mat"), {"class_names": random_class_names})
 
         return num_examples
 
diff --git a/torchvision/datasets/stanford_cars.py b/torchvision/datasets/stanford_cars.py
index 9af9dc3fb49..60f0c0df2d1 100644
--- a/torchvision/datasets/stanford_cars.py
+++ b/torchvision/datasets/stanford_cars.py
@@ -1,23 +1,26 @@
-import os
-import os.path
+import pathlib
 from typing import Callable, Optional, Any, Tuple
 
 from PIL import Image
 
-from .utils import download_and_extract_archive, download_url
+from .utils import download_and_extract_archive, download_url, verify_str_arg
 from .vision import VisionDataset
 
 
 class StanfordCars(VisionDataset):
     """`Stanford Cars <https://ai.stanford.edu/~jkrause/cars/car_dataset.html>`_ Dataset
 
-    .. warning::
+    The Cars dataset contains 16,185 images of 196 classes of cars. The data is
+    split into 8,144 training images and 8,041 testing images, where each class
+    has been split roughly in a 50-50 split
+
+    .. note::
 
         This class needs `scipy <https://docs.scipy.org/doc/>`_ to load target files from `.mat` format.
 
     Args:
         root (string): Root directory of dataset
-        train (bool, optional):If True, creates dataset from training set, otherwise creates from test set
+        split (string, optional): The dataset split, supports ``"train"`` (default) or ``"test"``.
         transform (callable, optional): A function/transform that  takes in an PIL image
             and returns a transformed version. E.g, ``transforms.RandomCrop``
         target_transform (callable, optional): A function/transform that takes in the
@@ -26,30 +29,10 @@ class StanfordCars(VisionDataset):
             puts it in root directory. If dataset is already downloaded, it is not
             downloaded again."""
 
-    urls = (
-        "https://ai.stanford.edu/~jkrause/car196/cars_test.tgz",
-        "https://ai.stanford.edu/~jkrause/car196/cars_train.tgz",
-    )  # test and train image urls
-
-    md5s = (
-        "4ce7ebf6a94d07f1952d94dd34c4d501",
-        "065e5b463ae28d29e77c1b4b166cfe61",
-    )  # md5checksum for test and train data
-
-    annot_urls = (
-        "https://ai.stanford.edu/~jkrause/car196/cars_test_annos_withlabels.mat",
-        "https://ai.stanford.edu/~jkrause/cars/car_devkit.tgz",
-    )  # annotations and labels for test and train
-
-    annot_md5s = (
-        "b0a2b23655a3edd16d84508592a98d10",
-        "c3b158d763b6e2245038c8ad08e45376",
-    )  # md5 checksum for annotations
-
     def __init__(
         self,
         root: str,
-        train: bool = True,
+        split: str = "train",
         transform: Optional[Callable] = None,
         target_transform: Optional[Callable] = None,
         download: bool = False,
@@ -62,7 +45,16 @@ def __init__(
 
         super().__init__(root, transform=transform, target_transform=target_transform)
 
-        self.train = train
+        self._split = verify_str_arg(split, "split", ("train", "test"))
+        self._base_folder = pathlib.Path(root) / "stanford_cars"
+        devkit = self._base_folder / "devkit"
+
+        if self._split == "train":
+            self._annotations_mat_path = devkit / "cars_train_annos.mat"
+            self._images_base_path = self._base_folder / "cars_train"
+        else:
+            self._annotations_mat_path = self._base_folder / "cars_test_annos_withlabels.mat"
+            self._images_base_path = self._base_folder / "cars_test"
 
         if download:
             self.download()
@@ -72,22 +64,13 @@ def __init__(
 
         self._samples = [
             (
-                os.path.join(self.root, f"cars_{'train' if self.train else 'test'}", annotation["fname"]),
-                annotation["class"] - 1,
-                # Beware stanford cars target mapping  starts from 1
+                str(self._images_base_path / annotation["fname"]),
+                annotation["class"] - 1,  # Original target mapping  starts from 1, hence -1
             )
-            for annotation in sio.loadmat(
-                os.path.join(
-                    self.root,
-                    *["devkit", "cars_train_annos.mat"] if self.train else ["cars_test_annos_withlabels.mat"],
-                ),
-                squeeze_me=True,
-            )["annotations"]
+            for annotation in sio.loadmat(self._annotations_mat_path, squeeze_me=True)["annotations"]
         ]
 
-        self.classes = sio.loadmat(os.path.join(self.root, "devkit", "cars_meta.mat"), squeeze_me=True)[
-            "class_names"
-        ].tolist()
+        self.classes = sio.loadmat(str(devkit / "cars_meta.mat"), squeeze_me=True)["class_names"].tolist()
         self.class_to_idx = {cls: i for i, cls in enumerate(self.classes)}
 
     def __len__(self) -> int:
@@ -108,20 +91,31 @@ def download(self) -> None:
         if self._check_exists():
             return
 
-        download_and_extract_archive(url=self.urls[self.train], download_root=self.root, md5=self.md5s[self.train])
-        download_and_extract_archive(url=self.annot_urls[1], download_root=self.root, md5=self.annot_md5s[1])
-        if not self.train:
+        download_and_extract_archive(
+            url="https://ai.stanford.edu/~jkrause/cars/car_devkit.tgz",
+            download_root=self._base_folder,
+            md5="c3b158d763b6e2245038c8ad08e45376",
+        )
+        if self._split == "train":
+            download_and_extract_archive(
+                url="https://ai.stanford.edu/~jkrause/car196/cars_train.tgz",
+                download_root=self._base_folder,
+                md5="065e5b463ae28d29e77c1b4b166cfe61",
+            )
+        else:
+            download_and_extract_archive(
+                url="https://ai.stanford.edu/~jkrause/car196/cars_test.tgz",
+                download_root=self._base_folder,
+                md5="4ce7ebf6a94d07f1952d94dd34c4d501",
+            )
             download_url(
-                url=self.annot_urls[0],
-                root=self.root,
-                md5=self.annot_md5s[0],
+                url="https://ai.stanford.edu/~jkrause/car196/cars_test_annos_withlabels.mat",
+                root=self._base_folder,
+                md5="b0a2b23655a3edd16d84508592a98d10",
             )
 
     def _check_exists(self) -> bool:
-        return (
-            os.path.exists(os.path.join(self.root, f"cars_{'train' if self.train else 'test'}"))
-            and os.path.isdir(os.path.join(self.root, f"cars_{'train' if self.train else 'test'}"))
-            and os.path.exists(os.path.join(self.root, "devkit", "cars_meta.mat"))
-            if self.train
-            else os.path.exists(os.path.join(self.root, "cars_test_annos_withlabels.mat"))
-        )
+        if not (self._base_folder / "devkit").is_dir():
+            return False
+
+        return self._annotations_mat_path.exists() and self._images_base_path.is_dir()

From dc463d767b850eb4f3a00f8faf4b4399cd2bbd3c Mon Sep 17 00:00:00 2001
From: Nicolas Hug <nicolashug@fb.com>
Date: Mon, 17 Jan 2022 18:19:13 +0000
Subject: [PATCH 14/15] Sigh

---
 torchvision/datasets/stanford_cars.py | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/torchvision/datasets/stanford_cars.py b/torchvision/datasets/stanford_cars.py
index 60f0c0df2d1..daca0b0b46a 100644
--- a/torchvision/datasets/stanford_cars.py
+++ b/torchvision/datasets/stanford_cars.py
@@ -93,24 +93,24 @@ def download(self) -> None:
 
         download_and_extract_archive(
             url="https://ai.stanford.edu/~jkrause/cars/car_devkit.tgz",
-            download_root=self._base_folder,
+            download_root=str(self._base_folder),
             md5="c3b158d763b6e2245038c8ad08e45376",
         )
         if self._split == "train":
             download_and_extract_archive(
                 url="https://ai.stanford.edu/~jkrause/car196/cars_train.tgz",
-                download_root=self._base_folder,
+                download_root=str(self._base_folder),
                 md5="065e5b463ae28d29e77c1b4b166cfe61",
             )
         else:
             download_and_extract_archive(
                 url="https://ai.stanford.edu/~jkrause/car196/cars_test.tgz",
-                download_root=self._base_folder,
+                download_root=str(self._base_folder),
                 md5="4ce7ebf6a94d07f1952d94dd34c4d501",
             )
             download_url(
                 url="https://ai.stanford.edu/~jkrause/car196/cars_test_annos_withlabels.mat",
-                root=self._base_folder,
+                root=str(self._base_folder),
                 md5="b0a2b23655a3edd16d84508592a98d10",
             )
 

From b593760a0eece516fb10b58eec7b0740b4e605b8 Mon Sep 17 00:00:00 2001
From: Nicolas Hug <nicolashug@fb.com>
Date: Tue, 18 Jan 2022 09:54:52 +0000
Subject: [PATCH 15/15] don't convert to strings

---
 test/test_datasets.py | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/test/test_datasets.py b/test/test_datasets.py
index a0ba37d49e8..b7e3d8e55fb 100644
--- a/test/test_datasets.py
+++ b/test/test_datasets.py
@@ -2532,10 +2532,10 @@ def inject_fake_data(self, tmpdir, config):
 
         if config["split"] == "train":
             images_folder_name = "cars_train"
-            annotations_mat_path = str(devkit / "cars_train_annos.mat")
+            annotations_mat_path = devkit / "cars_train_annos.mat"
         else:
             images_folder_name = "cars_test"
-            annotations_mat_path = str(base_folder / "cars_test_annos_withlabels.mat")
+            annotations_mat_path = base_folder / "cars_test_annos_withlabels.mat"
 
         datasets_utils.create_image_folder(
             root=base_folder,
@@ -2553,7 +2553,7 @@ def inject_fake_data(self, tmpdir, config):
         io.savemat(annotations_mat_path, {"annotations": rec_array})
 
         random_class_names = ["random_name"] * num_classes
-        io.savemat(str(devkit / "cars_meta.mat"), {"class_names": random_class_names})
+        io.savemat(devkit / "cars_meta.mat", {"class_names": random_class_names})
 
         return num_examples