fix: load script

dogs-video-object-tracking-dataset.py

@@ -39,8 +39,6 @@ class DogsVideoObjectTrackingDataset(datasets.GeneratorBasedBuilder):
 
     DEFAULT_CONFIG_NAME = "video_01"
 
-    DEFAULT_WRITER_BATCH_SIZE = 400000000
-
     def _info(self):
         return datasets.DatasetInfo(
             description=_DESCRIPTION,
@@ -101,7 +99,6 @@ class DogsVideoObjectTrackingDataset(datasets.GeneratorBasedBuilder):
         for track in root.iter("track"):
             shape = track.find(f".//*[@frame='{index}']")
             if not (shape is None):
-
                 shape.set("label", track.get("label"))
                 shape.set("track_id", track.get("id"))
                 img.append(shape)
@@ -153,7 +150,7 @@ class DogsVideoObjectTrackingDataset(datasets.GeneratorBasedBuilder):
         return shape_data
 
     def _generate_examples(self, data):
-        tree = ET.parse(f"{data}/annotations.xml")
+        tree = ET.parse(os.path.join(data, "annotations.xml"))
         root = tree.getroot()
 
         for idx, file in enumerate(sorted(os.listdir(f"{data}/images"))):