ultralytics 8.3.190__py3-none-any.whl → 8.3.191__py3-none-any.whl

ultralytics/data/base.py

@@ -1,5 +1,7 @@
 # Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license
 
+from __future__ import annotations
+
 import glob
 import math
 import os
@@ -7,7 +9,7 @@ import random
 from copy import deepcopy
 from multiprocessing.pool import ThreadPool
 from pathlib import Path
-from typing import Any, Dict, List, Optional, Tuple, Union
+from typing import Any
 
 import cv2
 import numpy as np
@@ -69,18 +71,18 @@ class BaseDataset(Dataset):
 
     def __init__(
         self,
-        img_path: Union[str, List[str]],
+        img_path: str | list[str],
         imgsz: int = 640,
-        cache: Union[bool, str] = False,
+        cache: bool | str = False,
         augment: bool = True,
-        hyp: Dict[str, Any] = DEFAULT_CFG,
+        hyp: dict[str, Any] = DEFAULT_CFG,
         prefix: str = "",
         rect: bool = False,
         batch_size: int = 16,
         stride: int = 32,
         pad: float = 0.5,
         single_cls: bool = False,
-        classes: Optional[List[int]] = None,
+        classes: list[int] | None = None,
         fraction: float = 1.0,
         channels: int = 3,
     ):
@@ -145,7 +147,7 @@ class BaseDataset(Dataset):
         # Transforms
         self.transforms = self.build_transforms(hyp=hyp)
 
-    def get_img_files(self, img_path: Union[str, List[str]]) -> List[str]:
+    def get_img_files(self, img_path: str | list[str]) -> list[str]:
         """
         Read image files from the specified path.
 
@@ -183,7 +185,7 @@ class BaseDataset(Dataset):
         check_file_speeds(im_files, prefix=self.prefix)  # check image read speeds
         return im_files
 
-    def update_labels(self, include_class: Optional[List[int]]) -> None:
+    def update_labels(self, include_class: list[int] | None) -> None:
         """
         Update labels to include only specified classes.
 
@@ -207,7 +209,7 @@ class BaseDataset(Dataset):
             if self.single_cls:
                 self.labels[i]["cls"][:, 0] = 0
 
-    def load_image(self, i: int, rect_mode: bool = True) -> Tuple[np.ndarray, Tuple[int, int], Tuple[int, int]]:
+    def load_image(self, i: int, rect_mode: bool = True) -> tuple[np.ndarray, tuple[int, int], tuple[int, int]]:
         """
         Load an image from dataset index 'i'.
 
@@ -374,11 +376,11 @@ class BaseDataset(Dataset):
         self.batch_shapes = np.ceil(np.array(shapes) * self.imgsz / self.stride + self.pad).astype(int) * self.stride
         self.batch = bi  # batch index of image
 
-    def __getitem__(self, index: int) -> Dict[str, Any]:
+    def __getitem__(self, index: int) -> dict[str, Any]:
         """Return transformed label information for given index."""
         return self.transforms(self.get_image_and_label(index))
 
-    def get_image_and_label(self, index: int) -> Dict[str, Any]:
+    def get_image_and_label(self, index: int) -> dict[str, Any]:
         """
         Get and return label information from the dataset.
 
@@ -403,11 +405,11 @@ class BaseDataset(Dataset):
         """Return the length of the labels list for the dataset."""
         return len(self.labels)
 
-    def update_labels_info(self, label: Dict[str, Any]) -> Dict[str, Any]:
+    def update_labels_info(self, label: dict[str, Any]) -> dict[str, Any]:
         """Custom your label format here."""
         return label
 
-    def build_transforms(self, hyp: Optional[Dict[str, Any]] = None):
+    def build_transforms(self, hyp: dict[str, Any] | None = None):
         """
         Users can customize augmentations here.
 
@@ -421,7 +423,7 @@ class BaseDataset(Dataset):
         """
         raise NotImplementedError
 
-    def get_labels(self) -> List[Dict[str, Any]]:
+    def get_labels(self) -> list[dict[str, Any]]:
         """
         Users can customize their own format here.
 

ultralytics/data/build.py

@@ -1,9 +1,12 @@
 # Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license
 
+from __future__ import annotations
+
 import os
 import random
+from collections.abc import Iterator
 from pathlib import Path
-from typing import Any, Dict, Iterator
+from typing import Any
 
 import numpy as np
 import torch
@@ -116,7 +119,7 @@ def build_yolo_dataset(
     cfg: IterableSimpleNamespace,
     img_path: str,
     batch: int,
-    data: Dict[str, Any],
+    data: dict[str, Any],
     mode: str = "train",
     rect: bool = False,
     stride: int = 32,
@@ -133,7 +136,7 @@ def build_yolo_dataset(
         rect=cfg.rect or rect,  # rectangular batches
         cache=cfg.cache or None,
         single_cls=cfg.single_cls or False,
-        stride=int(stride),
+        stride=stride,
         pad=0.0 if mode == "train" else 0.5,
         prefix=colorstr(f"{mode}: "),
         task=cfg.task,
@@ -165,7 +168,7 @@ def build_grounding(
         rect=cfg.rect or rect,  # rectangular batches
         cache=cfg.cache or None,
         single_cls=cfg.single_cls or False,
-        stride=int(stride),
+        stride=stride,
         pad=0.0 if mode == "train" else 0.5,
         prefix=colorstr(f"{mode}: "),
         task=cfg.task,

ultralytics/data/converter.py

@@ -1,5 +1,7 @@
 # Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license
 
+from __future__ import annotations
+
 import asyncio
 import json
 import random
@@ -7,7 +9,6 @@ import shutil
 from collections import defaultdict
 from concurrent.futures import ThreadPoolExecutor, as_completed
 from pathlib import Path
-from typing import List, Optional, Union
 
 import cv2
 import numpy as np
@@ -19,7 +20,7 @@ from ultralytics.utils.downloads import download, zip_directory
 from ultralytics.utils.files import increment_path
 
 
-def coco91_to_coco80_class() -> List[int]:
+def coco91_to_coco80_class() -> list[int]:
     """
     Convert 91-index COCO class IDs to 80-index COCO class IDs.
 
@@ -122,7 +123,7 @@ def coco91_to_coco80_class() -> List[int]:
     ]
 
 
-def coco80_to_coco91_class() -> List[int]:
+def coco80_to_coco91_class() -> list[int]:
     r"""
     Convert 80-index (val2014) to 91-index (paper).
 
@@ -531,7 +532,7 @@ def min_index(arr1: np.ndarray, arr2: np.ndarray):
     return np.unravel_index(np.argmin(dis, axis=None), dis.shape)
 
 
-def merge_multi_segment(segments: List[List]):
+def merge_multi_segment(segments: list[list]):
     """
     Merge multiple segments into one list by connecting the coordinates with the minimum distance between each segment.
 
@@ -582,9 +583,7 @@ def merge_multi_segment(segments: List[List]):
     return s
 
 
-def yolo_bbox2segment(
-    im_dir: Union[str, Path], save_dir: Optional[Union[str, Path]] = None, sam_model: str = "sam_b.pt", device=None
-):
+def yolo_bbox2segment(im_dir: str | Path, save_dir: str | Path | None = None, sam_model: str = "sam_b.pt", device=None):
     """
     Convert existing object detection dataset (bounding boxes) to segmentation dataset or oriented bounding box (OBB) in
     YOLO format. Generate segmentation data using SAM auto-annotator as needed.
@@ -706,7 +705,7 @@ def create_synthetic_coco_dataset():
     LOGGER.info("Synthetic COCO dataset created successfully.")
 
 
-def convert_to_multispectral(path: Union[str, Path], n_channels: int = 10, replace: bool = False, zip: bool = False):
+def convert_to_multispectral(path: str | Path, n_channels: int = 10, replace: bool = False, zip: bool = False):
     """
     Convert RGB images to multispectral images by interpolating across wavelength bands.
 
@@ -733,7 +732,7 @@ def convert_to_multispectral(path: Union[str, Path], n_channels: int = 10, repla
     path = Path(path)
     if path.is_dir():
         # Process directory
-        im_files = sum([list(path.rglob(f"*.{ext}")) for ext in (IMG_FORMATS - {"tif", "tiff"})], [])
+        im_files = sum((list(path.rglob(f"*.{ext}")) for ext in (IMG_FORMATS - {"tif", "tiff"})), [])
         for im_path in im_files:
             try:
                 convert_to_multispectral(im_path, n_channels)
@@ -758,7 +757,7 @@ def convert_to_multispectral(path: Union[str, Path], n_channels: int = 10, repla
         LOGGER.info(f"Converted {output_path}")
 
 
-async def convert_ndjson_to_yolo(ndjson_path: Union[str, Path], output_path: Optional[Union[str, Path]] = None) -> Path:
+async def convert_ndjson_to_yolo(ndjson_path: str | Path, output_path: str | Path | None = None) -> Path:
     """
     Convert NDJSON dataset format to Ultralytics YOLO11 dataset structure.
 

ultralytics/data/dataset.py

@@ -1,11 +1,13 @@
 # Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license
 
+from __future__ import annotations
+
 import json
 from collections import defaultdict
 from itertools import repeat
 from multiprocessing.pool import ThreadPool
 from pathlib import Path
-from typing import Any, Dict, List, Optional, Tuple
+from typing import Any
 
 import cv2
 import numpy as np
@@ -70,7 +72,7 @@ class YOLODataset(BaseDataset):
         >>> dataset.get_labels()
     """
 
-    def __init__(self, *args, data: Optional[Dict] = None, task: str = "detect", **kwargs):
+    def __init__(self, *args, data: dict | None = None, task: str = "detect", **kwargs):
         """
         Initialize the YOLODataset.
 
@@ -87,7 +89,7 @@ class YOLODataset(BaseDataset):
         assert not (self.use_segments and self.use_keypoints), "Can not use both segments and keypoints."
         super().__init__(*args, channels=self.data.get("channels", 3), **kwargs)
 
-    def cache_labels(self, path: Path = Path("./labels.cache")) -> Dict:
+    def cache_labels(self, path: Path = Path("./labels.cache")) -> dict:
         """
         Cache dataset labels, check images and read shapes.
 
@@ -155,7 +157,7 @@ class YOLODataset(BaseDataset):
         save_dataset_cache_file(self.prefix, path, x, DATASET_CACHE_VERSION)
         return x
 
-    def get_labels(self) -> List[Dict]:
+    def get_labels(self) -> list[dict]:
         """
         Return dictionary of labels for YOLO training.
 
@@ -205,7 +207,7 @@ class YOLODataset(BaseDataset):
             LOGGER.warning(f"Labels are missing or empty in {cache_path}, training may not work correctly. {HELP_URL}")
         return labels
 
-    def build_transforms(self, hyp: Optional[Dict] = None) -> Compose:
+    def build_transforms(self, hyp: dict | None = None) -> Compose:
         """
         Build and append transforms to the list.
 
@@ -237,7 +239,7 @@ class YOLODataset(BaseDataset):
         )
         return transforms
 
-    def close_mosaic(self, hyp: Dict) -> None:
+    def close_mosaic(self, hyp: dict) -> None:
         """
         Disable mosaic, copy_paste, mixup and cutmix augmentations by setting their probabilities to 0.0.
 
@@ -250,7 +252,7 @@ class YOLODataset(BaseDataset):
         hyp.cutmix = 0.0
         self.transforms = self.build_transforms(hyp)
 
-    def update_labels_info(self, label: Dict) -> Dict:
+    def update_labels_info(self, label: dict) -> dict:
         """
         Update label format for different tasks.
 
@@ -284,7 +286,7 @@ class YOLODataset(BaseDataset):
         return label
 
     @staticmethod
-    def collate_fn(batch: List[Dict]) -> Dict:
+    def collate_fn(batch: list[dict]) -> dict:
         """
         Collate data samples into batches.
 
@@ -331,7 +333,7 @@ class YOLOMultiModalDataset(YOLODataset):
         >>> print(batch.keys())  # Should include 'texts'
     """
 
-    def __init__(self, *args, data: Optional[Dict] = None, task: str = "detect", **kwargs):
+    def __init__(self, *args, data: dict | None = None, task: str = "detect", **kwargs):
         """
         Initialize a YOLOMultiModalDataset.
 
@@ -343,7 +345,7 @@ class YOLOMultiModalDataset(YOLODataset):
         """
         super().__init__(*args, data=data, task=task, **kwargs)
 
-    def update_labels_info(self, label: Dict) -> Dict:
+    def update_labels_info(self, label: dict) -> dict:
         """
         Add text information for multi-modal model training.
 
@@ -360,7 +362,7 @@ class YOLOMultiModalDataset(YOLODataset):
 
         return labels
 
-    def build_transforms(self, hyp: Optional[Dict] = None) -> Compose:
+    def build_transforms(self, hyp: dict | None = None) -> Compose:
         """
         Enhance data transformations with optional text augmentation for multi-modal training.
 
@@ -409,7 +411,7 @@ class YOLOMultiModalDataset(YOLODataset):
         return category_freq
 
     @staticmethod
-    def _get_neg_texts(category_freq: Dict, threshold: int = 100) -> List[str]:
+    def _get_neg_texts(category_freq: dict, threshold: int = 100) -> list[str]:
         """Get negative text samples based on frequency threshold."""
         threshold = min(max(category_freq.values()), 100)
         return [k for k, v in category_freq.items() if v >= threshold]
@@ -451,7 +453,7 @@ class GroundingDataset(YOLODataset):
         self.max_samples = max_samples
         super().__init__(*args, task=task, data={"channels": 3}, **kwargs)
 
-    def get_img_files(self, img_path: str) -> List:
+    def get_img_files(self, img_path: str) -> list:
         """
         The image files would be read in `get_labels` function, return empty list here.
 
@@ -463,7 +465,7 @@ class GroundingDataset(YOLODataset):
         """
         return []
 
-    def verify_labels(self, labels: List[Dict[str, Any]]) -> None:
+    def verify_labels(self, labels: list[dict[str, Any]]) -> None:
         """
         Verify the number of instances in the dataset matches expected counts.
 
@@ -498,7 +500,7 @@ class GroundingDataset(YOLODataset):
                 return
         LOGGER.warning(f"Skipping instance count verification for unrecognized dataset '{self.json_file}'")
 
-    def cache_labels(self, path: Path = Path("./labels.cache")) -> Dict[str, Any]:
+    def cache_labels(self, path: Path = Path("./labels.cache")) -> dict[str, Any]:
         """
         Load annotations from a JSON file, filter, and normalize bounding boxes for each image.
 
@@ -589,7 +591,7 @@ class GroundingDataset(YOLODataset):
         save_dataset_cache_file(self.prefix, path, x, DATASET_CACHE_VERSION)
         return x
 
-    def get_labels(self) -> List[Dict]:
+    def get_labels(self) -> list[dict]:
         """
         Load labels from cache or generate them from JSON file.
 
@@ -611,7 +613,7 @@ class GroundingDataset(YOLODataset):
             LOGGER.info(f"Load {self.json_file} from cache file {cache_path}")
         return labels
 
-    def build_transforms(self, hyp: Optional[Dict] = None) -> Compose:
+    def build_transforms(self, hyp: dict | None = None) -> Compose:
         """
         Configure augmentations for training with optional text loading.
 
@@ -652,7 +654,7 @@ class GroundingDataset(YOLODataset):
         return category_freq
 
     @staticmethod
-    def _get_neg_texts(category_freq: Dict, threshold: int = 100) -> List[str]:
+    def _get_neg_texts(category_freq: dict, threshold: int = 100) -> list[str]:
         """Get negative text samples based on frequency threshold."""
         threshold = min(max(category_freq.values()), 100)
         return [k for k, v in category_freq.items() if v >= threshold]
@@ -675,7 +677,7 @@ class YOLOConcatDataset(ConcatDataset):
     """
 
     @staticmethod
-    def collate_fn(batch: List[Dict]) -> Dict:
+    def collate_fn(batch: list[dict]) -> dict:
         """
         Collate data samples into batches.
 
@@ -687,7 +689,7 @@ class YOLOConcatDataset(ConcatDataset):
         """
         return YOLODataset.collate_fn(batch)
 
-    def close_mosaic(self, hyp: Dict) -> None:
+    def close_mosaic(self, hyp: dict) -> None:
         """
         Set mosaic, copy_paste and mixup options to 0.0 and build transformations.
 
@@ -784,7 +786,7 @@ class ClassificationDataset:
             else classify_transforms(size=args.imgsz)
         )
 
-    def __getitem__(self, i: int) -> Dict:
+    def __getitem__(self, i: int) -> dict:
         """
         Return subset of data and targets corresponding to given indices.
 
@@ -813,7 +815,7 @@ class ClassificationDataset:
         """Return the total number of samples in the dataset."""
         return len(self.samples)
 
-    def verify_images(self) -> List[Tuple]:
+    def verify_images(self) -> list[tuple]:
         """
         Verify all images in dataset.
 

ultralytics/data/loaders.py

@@ -1,5 +1,7 @@
 # Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license
 
+from __future__ import annotations
+
 import glob
 import math
 import os
@@ -8,7 +10,7 @@ import urllib
 from dataclasses import dataclass
 from pathlib import Path
 from threading import Thread
-from typing import Any, List, Optional, Tuple, Union
+from typing import Any
 
 import cv2
 import numpy as np
@@ -192,7 +194,7 @@ class LoadStreams:
         self.count = -1
         return self
 
-    def __next__(self) -> Tuple[List[str], List[np.ndarray], List[str]]:
+    def __next__(self) -> tuple[list[str], list[np.ndarray], list[str]]:
         """Return the next batch of frames from multiple video streams for processing."""
         self.count += 1
 
@@ -294,7 +296,7 @@ class LoadScreenshots:
         """Yield the next screenshot image from the specified screen or region for processing."""
         return self
 
-    def __next__(self) -> Tuple[List[str], List[np.ndarray], List[str]]:
+    def __next__(self) -> tuple[list[str], list[np.ndarray], list[str]]:
         """Capture and return the next screenshot as a numpy array using the mss library."""
         im0 = np.asarray(self.sct.grab(self.monitor))[:, :, :3]  # BGRA to BGR
         im0 = cv2.cvtColor(im0, cv2.COLOR_BGR2GRAY)[..., None] if self.cv2_flag == cv2.IMREAD_GRAYSCALE else im0
@@ -344,7 +346,7 @@ class LoadImagesAndVideos:
         - Can read from a text file containing paths to images and videos.
     """
 
-    def __init__(self, path: Union[str, Path, List], batch: int = 1, vid_stride: int = 1, channels: int = 3):
+    def __init__(self, path: str | Path | list, batch: int = 1, vid_stride: int = 1, channels: int = 3):
         """
         Initialize dataloader for images and videos, supporting various input formats.
 
@@ -403,7 +405,7 @@ class LoadImagesAndVideos:
         self.count = 0
         return self
 
-    def __next__(self) -> Tuple[List[str], List[np.ndarray], List[str]]:
+    def __next__(self) -> tuple[list[str], list[np.ndarray], list[str]]:
         """Return the next batch of images or video frames with their paths and metadata."""
         paths, imgs, info = [], [], []
         while len(imgs) < self.bs:
@@ -514,7 +516,7 @@ class LoadPilAndNumpy:
         Loaded 2 images
     """
 
-    def __init__(self, im0: Union[Image.Image, np.ndarray, List], channels: int = 3):
+    def __init__(self, im0: Image.Image | np.ndarray | list, channels: int = 3):
         """
         Initialize a loader for PIL and Numpy images, converting inputs to a standardized format.
 
@@ -532,7 +534,7 @@ class LoadPilAndNumpy:
         self.bs = len(self.im0)
 
     @staticmethod
-    def _single_check(im: Union[Image.Image, np.ndarray], flag: str = "RGB") -> np.ndarray:
+    def _single_check(im: Image.Image | np.ndarray, flag: str = "RGB") -> np.ndarray:
         """Validate and format an image to numpy array, ensuring RGB order and contiguous memory."""
         assert isinstance(im, (Image.Image, np.ndarray)), f"Expected PIL/np.ndarray image type, but got {type(im)}"
         if isinstance(im, Image.Image):
@@ -548,7 +550,7 @@ class LoadPilAndNumpy:
         """Return the length of the 'im0' attribute, representing the number of loaded images."""
         return len(self.im0)
 
-    def __next__(self) -> Tuple[List[str], List[np.ndarray], List[str]]:
+    def __next__(self) -> tuple[list[str], list[np.ndarray], list[str]]:
         """Return the next batch of images, paths, and metadata for processing."""
         if self.count == 1:  # loop only once as it's batch inference
             raise StopIteration
@@ -624,7 +626,7 @@ class LoadTensor:
         self.count = 0
         return self
 
-    def __next__(self) -> Tuple[List[str], torch.Tensor, List[str]]:
+    def __next__(self) -> tuple[list[str], torch.Tensor, list[str]]:
         """Yield the next batch of tensor images and metadata for processing."""
         if self.count == 1:
             raise StopIteration
@@ -636,7 +638,7 @@ class LoadTensor:
         return self.bs
 
 
-def autocast_list(source: List[Any]) -> List[Union[Image.Image, np.ndarray]]:
+def autocast_list(source: list[Any]) -> list[Image.Image | np.ndarray]:
     """Merge a list of sources into a list of numpy arrays or PIL images for Ultralytics prediction."""
     files = []
     for im in source:
@@ -653,7 +655,7 @@ def autocast_list(source: List[Any]) -> List[Union[Image.Image, np.ndarray]]:
     return files
 
 
-def get_best_youtube_url(url: str, method: str = "pytube") -> Optional[str]:
+def get_best_youtube_url(url: str, method: str = "pytube") -> str | None:
     """
     Retrieve the URL of the best quality MP4 video stream from a given YouTube video.
 

ultralytics/data/split.py

@@ -1,15 +1,16 @@
 # Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license
 
+from __future__ import annotations
+
 import random
 import shutil
 from pathlib import Path
-from typing import Tuple, Union
 
 from ultralytics.data.utils import IMG_FORMATS, img2label_paths
 from ultralytics.utils import DATASETS_DIR, LOGGER, TQDM
 
 
-def split_classify_dataset(source_dir: Union[str, Path], train_ratio: float = 0.8) -> Path:
+def split_classify_dataset(source_dir: str | Path, train_ratio: float = 0.8) -> Path:
     """
     Split classification dataset into train and val directories in a new directory.
 
@@ -97,7 +98,7 @@ def split_classify_dataset(source_dir: Union[str, Path], train_ratio: float = 0.
 
 def autosplit(
     path: Path = DATASETS_DIR / "coco8/images",
-    weights: Tuple[float, float, float] = (0.9, 0.1, 0.0),
+    weights: tuple[float, float, float] = (0.9, 0.1, 0.0),
     annotated_only: bool = False,
 ) -> None:
     """

ultralytics/data/split_dota.py

@@ -1,10 +1,12 @@
 # Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license
 
+from __future__ import annotations
+
 import itertools
 from glob import glob
 from math import ceil
 from pathlib import Path
-from typing import Any, Dict, List, Tuple
+from typing import Any
 
 import cv2
 import numpy as np
@@ -62,7 +64,7 @@ def bbox_iof(polygon1: np.ndarray, bbox2: np.ndarray, eps: float = 1e-6) -> np.n
     return outputs
 
 
-def load_yolo_dota(data_root: str, split: str = "train") -> List[Dict[str, Any]]:
+def load_yolo_dota(data_root: str, split: str = "train") -> list[dict[str, Any]]:
     """
     Load DOTA dataset annotations and image information.
 
@@ -99,9 +101,9 @@ def load_yolo_dota(data_root: str, split: str = "train") -> List[Dict[str, Any]]
 
 
 def get_windows(
-    im_size: Tuple[int, int],
-    crop_sizes: Tuple[int, ...] = (1024,),
-    gaps: Tuple[int, ...] = (200,),
+    im_size: tuple[int, int],
+    crop_sizes: tuple[int, ...] = (1024,),
+    gaps: tuple[int, ...] = (200,),
     im_rate_thr: float = 0.6,
     eps: float = 0.01,
 ) -> np.ndarray:
@@ -151,7 +153,7 @@ def get_windows(
     return windows[im_rates > im_rate_thr]
 
 
-def get_window_obj(anno: Dict[str, Any], windows: np.ndarray, iof_thr: float = 0.7) -> List[np.ndarray]:
+def get_window_obj(anno: dict[str, Any], windows: np.ndarray, iof_thr: float = 0.7) -> list[np.ndarray]:
     """Get objects for each window based on IoF threshold."""
     h, w = anno["ori_size"]
     label = anno["label"]
@@ -166,9 +168,9 @@ def get_window_obj(anno: Dict[str, Any], windows: np.ndarray, iof_thr: float = 0
 
 
 def crop_and_save(
-    anno: Dict[str, Any],
+    anno: dict[str, Any],
     windows: np.ndarray,
-    window_objs: List[np.ndarray],
+    window_objs: list[np.ndarray],
     im_dir: str,
     lb_dir: str,
     allow_background_images: bool = True,
@@ -221,8 +223,8 @@ def split_images_and_labels(
     data_root: str,
     save_dir: str,
     split: str = "train",
-    crop_sizes: Tuple[int, ...] = (1024,),
-    gaps: Tuple[int, ...] = (200,),
+    crop_sizes: tuple[int, ...] = (1024,),
+    gaps: tuple[int, ...] = (200,),
 ) -> None:
     """
     Split both images and labels for a given dataset split.
@@ -261,7 +263,7 @@ def split_images_and_labels(
 
 
 def split_trainval(
-    data_root: str, save_dir: str, crop_size: int = 1024, gap: int = 200, rates: Tuple[float, ...] = (1.0,)
+    data_root: str, save_dir: str, crop_size: int = 1024, gap: int = 200, rates: tuple[float, ...] = (1.0,)
 ) -> None:
     """
     Split train and val sets of DOTA dataset with multiple scaling rates.
@@ -300,7 +302,7 @@ def split_trainval(
 
 
 def split_test(
-    data_root: str, save_dir: str, crop_size: int = 1024, gap: int = 200, rates: Tuple[float, ...] = (1.0,)
+    data_root: str, save_dir: str, crop_size: int = 1024, gap: int = 200, rates: tuple[float, ...] = (1.0,)
 ) -> None:
     """
     Split test set of DOTA dataset, labels are not included within this set.