fix

README.md

@@ -45,10 +45,10 @@ dataset_info:
         dtype: string
   splits:
   - name: train
-    num_bytes: 25910
+    num_bytes: 18266
     num_examples: 52
-  download_size: 313334253
-  dataset_size: 25910
+  download_size: 474311303
+  dataset_size: 18266
 - config_name: video_02
   features:
   - name: id
@@ -85,10 +85,10 @@ dataset_info:
         dtype: string
   splits:
   - name: train
-    num_bytes: 28900
+    num_bytes: 20374
     num_examples: 58
-  download_size: 67368145
-  dataset_size: 28900
+  download_size: 101046098
+  dataset_size: 20374
 - config_name: video_03
   features:
   - name: id
@@ -125,10 +125,10 @@ dataset_info:
         dtype: string
   splits:
   - name: train
-    num_bytes: 24416
+    num_bytes: 17213
     num_examples: 49
-  download_size: 148418007
-  dataset_size: 24416
+  download_size: 224796300
+  dataset_size: 17213
 ---
 
 # Dogs Video Object Tracking Dataset

dogs-video-object-tracking-dataset.py

@@ -79,14 +79,16 @@ class DogsVideoObjectTrackingDataset(datasets.GeneratorBasedBuilder):
         )
 
     def _split_generators(self, dl_manager):
-        data = dl_manager.download_and_extract(f"{_DATA}{self.config.name}.zip")
+        images = dl_manager.download_and_extract(f"{_DATA}{self.config.name}.zip")
+        masks = dl_manager.download_and_extract(f"{_DATA}{self.config.name}_masks.zip")
         annotations = dl_manager.download(f"{_DATA}{self.config.name}.xml")
-        data = dl_manager.iter_files(data)
+        images = dl_manager.iter_files(images)
         return [
             datasets.SplitGenerator(
                 name=datasets.Split.TRAIN,
                 gen_kwargs={
-                    "data": data,
+                    "images": images,
+                    "masks": masks,
                     "annotations": annotations,
                 },
             ),
@@ -152,14 +154,14 @@ class DogsVideoObjectTrackingDataset(datasets.GeneratorBasedBuilder):
 
         return shape_data
 
-    def _generate_examples(self, data, annotations):
+    def _generate_examples(self, images, masks, annotations):
         tree = ET.parse(annotations)
         root = tree.getroot()
 
-        for idx, image_path in enumerate(data):
-            if "images" in image_path:
-                i = image_path.split("_")[-1][:2]
-                image_name = image_path.split("/")[3]
+        for idx, (image, mask) in enumerate(zip(images, masks)):
+            if "images" in image:
+                i = image.split("_")[-1][:2]
+                image_name = image.split("/")[3]
                 img = self.extract_shapes_from_tracks(root, image_name, i)
 
                 image_id = img.get("id")
@@ -168,8 +170,8 @@ class DogsVideoObjectTrackingDataset(datasets.GeneratorBasedBuilder):
 
                 yield idx, {
                     "id": image_id,
-                    "name": image_path,
-                    "image": image_path,
-                    "mask": f"{image_path.replace('images', 'boxes')}",
+                    "name": image,
+                    "image": image,
+                    "mask": mask,
                     "shapes": shapes,
                 }