ultralytics 8.0.72 faster Windows trainings and corrupt cache fix (#1912)

Co-authored-by: andreaswimmer <53872150+andreaswimmer@users.noreply.github.com>
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>

ultralytics/yolo/cfg/default.yaml

@@ -28,7 +28,7 @@ single_cls: False  # train multi-class data as single-class
 image_weights: False  # use weighted image selection for training
 rect: False  # rectangular training if mode='train' or rectangular validation if mode='val'
 cos_lr: False  # use cosine learning rate scheduler
-close_mosaic: 10  # disable mosaic augmentation for final 10 epochs
+close_mosaic: 0  # (int) disable mosaic augmentation for final epochs
 resume: False  # resume training from last checkpoint
 amp: True  # Automatic Mixed Precision (AMP) training, choices=[True, False], True runs AMP check
 # Segmentation

ultralytics/yolo/data/build.py

@@ -162,7 +162,18 @@ def check_source(source):
 
 def load_inference_source(source=None, transforms=None, imgsz=640, vid_stride=1, stride=32, auto=True):
     """
-    TODO: docs
+    Loads an inference source for object detection and applies necessary transformations.
+
+    Args:
+        source (str, Path, Tensor, PIL.Image, np.ndarray): The input source for inference.
+        transforms (callable, optional): Custom transformations to be applied to the input source.
+        imgsz (int, optional): The size of the image for inference. Default is 640.
+        vid_stride (int, optional): The frame interval for video sources. Default is 1.
+        stride (int, optional): The model stride. Default is 32.
+        auto (bool, optional): Automatically apply pre-processing. Default is True.
+
+    Returns:
+        dataset: A dataset object for the specified input source.
     """
     source, webcam, screenshot, from_img, in_memory, tensor = check_source(source)
     source_type = source.source_type if in_memory else SourceTypes(webcam, screenshot, from_img, tensor)
@@ -179,7 +190,6 @@ def load_inference_source(source=None, transforms=None, imgsz=640, vid_stride=1,
                               auto=auto,
                               transforms=transforms,
                               vid_stride=vid_stride)
-
     elif screenshot:
         dataset = LoadScreenshots(source, imgsz=imgsz, stride=stride, auto=auto, transforms=transforms)
     elif from_img:
@@ -192,6 +202,7 @@ def load_inference_source(source=None, transforms=None, imgsz=640, vid_stride=1,
                              transforms=transforms,
                              vid_stride=vid_stride)
 
-    setattr(dataset, 'source_type', source_type)  # attach source types
+    # Attach source types to the dataset
+    setattr(dataset, 'source_type', source_type)
 
     return dataset

ultralytics/yolo/data/dataset.py

@@ -77,7 +77,6 @@ class YOLODataset(BaseDataset):
         nm, nf, ne, nc, msgs = 0, 0, 0, 0, []  # number missing, found, empty, corrupt, messages
         desc = f'{self.prefix}Scanning {path.parent / path.stem}...'
         total = len(self.im_files)
-        nc = len(self.data['names'])
         nkpt, ndim = self.data.get('kpt_shape', (0, 0))
         if self.use_keypoints and (nkpt <= 0 or ndim not in (2, 3)):
             raise ValueError("'kpt_shape' in data.yaml missing or incorrect. Should be a list with [number of "

ultralytics/yolo/engine/model.py

@@ -253,10 +253,10 @@ class YOLO:
             source (str, optional): The input source for object tracking. Can be a file path or a video stream.
             stream (bool, optional): Whether the input source is a video stream. Defaults to False.
             persist (bool, optional): Whether to persist the trackers if they already exist. Defaults to False.
-            **kwargs: Additional keyword arguments for the tracking process.
+            **kwargs (optional): Additional keyword arguments for the tracking process.
 
         Returns:
-            object: The tracking results.
+            (List[ultralytics.yolo.engine.results.Results]): The tracking results.
 
         """
         if not hasattr(self.predictor, 'trackers'):

ultralytics/yolo/engine/results.py

@@ -244,8 +244,7 @@ class Boxes(BaseTensor):
         orig_shape (tuple): Original image size, in the format (height, width).
 
     Attributes:
-        boxes (torch.Tensor) or (numpy.ndarray): A tensor or numpy array containing the detection boxes,
-            with shape (num_boxes, 6).
+        boxes (torch.Tensor) or (numpy.ndarray): The detection boxes with shape (num_boxes, 6).
         orig_shape (torch.Tensor) or (numpy.ndarray): Original image size, in the format (height, width).
         is_track (bool): True if the boxes also include track IDs, False otherwise.
 
@@ -272,7 +271,6 @@ class Boxes(BaseTensor):
             boxes = boxes[None, :]
         n = boxes.shape[-1]
         assert n in (6, 7), f'expected `n` in [6, 7], but got {n}'  # xyxy, (track_id), conf, cls
-        # TODO
         self.is_track = n == 7
         self.boxes = boxes
         self.orig_shape = torch.as_tensor(orig_shape, device=boxes.device) if isinstance(boxes, torch.Tensor) \

ultralytics/yolo/utils/__init__.py

@@ -102,6 +102,7 @@ np.set_printoptions(linewidth=320, formatter={'float_kind': '{:11.5g}'.format})
 cv2.setNumThreads(0)  # prevent OpenCV from multithreading (incompatible with PyTorch DataLoader)
 os.environ['NUMEXPR_MAX_THREADS'] = str(NUM_THREADS)  # NumExpr max threads
 os.environ['CUBLAS_WORKSPACE_CONFIG'] = ':4096:8'  # for deterministic training
+os.environ['TF_CPP_MIN_LOG_LEVEL'] = '2'  # suppress verbose TF compiler warnings in Colab
 
 
 class SimpleClass: