dgenerate-ultralytics-headless 8.3.236__py3-none-any.whl → 8.3.239__py3-none-any.whl

ultralytics/models/sam/sam3/vl_combiner.py

@@ -0,0 +1,160 @@
+# Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license
+
+# Copyright (c) Meta Platforms, Inc. and affiliates. All Rights Reserved
+
+"""Provides utility to combine a vision backbone with a language backbone."""
+
+from __future__ import annotations
+
+from copy import copy
+
+import torch
+import torch.nn as nn
+from torch.nn.attention import SDPBackend, sdpa_kernel
+
+from .necks import Sam3DualViTDetNeck
+
+
+class SAM3VLBackbone(nn.Module):
+    """This backbone combines a vision backbone and a language backbone without fusion. As such it is more of a
+    convenience wrapper to handle the two backbones together.
+
+    It adds support for activation checkpointing and compilation.
+    """
+
+    def __init__(
+        self,
+        visual: Sam3DualViTDetNeck,
+        text,
+        compile_visual: bool = False,
+        act_ckpt_whole_vision_backbone: bool = False,
+        act_ckpt_whole_language_backbone: bool = False,
+        scalp=0,
+    ):
+        """Initialize the backbone combiner.
+
+        :param visual: The vision backbone to use
+        :param text: The text encoder to use
+        """
+        super().__init__()
+        self.vision_backbone: Sam3DualViTDetNeck = torch.compile(visual) if compile_visual else visual
+        self.language_backbone = text
+        self.scalp = scalp
+        # allow running activation checkpointing on the entire vision and language backbones
+        self.act_ckpt_whole_vision_backbone = act_ckpt_whole_vision_backbone
+        self.act_ckpt_whole_language_backbone = act_ckpt_whole_language_backbone
+
+    def forward(
+        self,
+        samples: torch.Tensor,
+        captions: list[str],
+        input_boxes: torch.Tensor = None,
+        additional_text: list[str] | None = None,
+    ):
+        """Forward pass of the backbone combiner.
+
+        :param samples: The input images
+        :param captions: The input captions
+        :param input_boxes: If the text contains place-holders for boxes, this
+            parameter contains the tensor containing their spatial features
+        :param additional_text: This can be used to encode some additional text
+            (different from the captions) in the same forward of the backbone
+        :return: Output dictionary with the following keys:
+            - vision_features: The output of the vision backbone
+            - language_features: The output of the language backbone
+            - language_mask: The attention mask of the language backbone
+            - vision_pos_enc: The positional encoding of the vision backbone
+            - (optional) additional_text_features: The output of the language
+                backbone for the additional text
+            - (optional) additional_text_mask: The attention mask of the
+                language backbone for the additional text
+        """
+        output = self.forward_image(samples)
+        output.update(self.forward_text(captions, input_boxes, additional_text))
+        return output
+
+    def forward_image(self, samples: torch.Tensor):
+        """Forward pass of the vision backbone and get both SAM3 and SAM2 features."""
+        # Forward through backbone
+        sam3_features, sam3_pos, sam2_features, sam2_pos = self.vision_backbone.forward(samples)
+        if self.scalp > 0:
+            # Discard the lowest resolution features
+            sam3_features, sam3_pos = (
+                sam3_features[: -self.scalp],
+                sam3_pos[: -self.scalp],
+            )
+            if sam2_features is not None and sam2_pos is not None:
+                sam2_features, sam2_pos = (
+                    sam2_features[: -self.scalp],
+                    sam2_pos[: -self.scalp],
+                )
+
+        sam2_output = None
+
+        if sam2_features is not None and sam2_pos is not None:
+            sam2_src = sam2_features[-1]
+            sam2_output = {
+                "vision_features": sam2_src,
+                "vision_pos_enc": sam2_pos,
+                "backbone_fpn": sam2_features,
+            }
+
+        sam3_src = sam3_features[-1]
+        return {
+            "vision_features": sam3_src,
+            "vision_pos_enc": sam3_pos,
+            "backbone_fpn": sam3_features,
+            "sam2_backbone_out": sam2_output,
+        }
+
+    def forward_image_sam2(self, samples: torch.Tensor):
+        """Forward pass of the vision backbone to get SAM2 features only."""
+        xs = self.vision_backbone.trunk(samples)
+        x = xs[-1]  # simpleFPN
+
+        assert self.vision_backbone.sam2_convs is not None, "SAM2 neck is not available."
+        sam2_features, sam2_pos = self.vision_backbone.sam_forward_feature_levels(x, self.vision_backbone.sam2_convs)
+
+        if self.scalp > 0:
+            # Discard the lowest resolution features
+            sam2_features, sam2_pos = (
+                sam2_features[: -self.scalp],
+                sam2_pos[: -self.scalp],
+            )
+
+        return {
+            "vision_features": sam2_features[-1],
+            "vision_pos_enc": sam2_pos,
+            "backbone_fpn": sam2_features,
+        }
+
+    def forward_text(self, captions, input_boxes=None, additional_text=None):
+        """Forward pass of the text encoder."""
+        output = {}
+
+        # Forward through text_encoder
+        text_to_encode = copy(captions)
+        if additional_text is not None:
+            # if there are additional_text, we piggy-back them into this forward.
+            # They'll be used later for output alignment
+            text_to_encode += additional_text
+
+        with sdpa_kernel([SDPBackend.MATH, SDPBackend.EFFICIENT_ATTENTION, SDPBackend.FLASH_ATTENTION]):
+            text_attention_mask, text_memory, text_embeds = self.language_backbone(text_to_encode, input_boxes)
+
+        if additional_text is not None:
+            output["additional_text_features"] = text_memory[:, -len(additional_text) :]
+            output["additional_text_mask"] = text_attention_mask[-len(additional_text) :]
+
+        text_memory = text_memory[:, : len(captions)]
+        text_attention_mask = text_attention_mask[: len(captions)]
+        text_embeds = text_embeds[:, : len(captions)]
+        output["language_features"] = text_memory
+        output["language_mask"] = text_attention_mask
+        output["language_embeds"] = text_embeds  # Text embeddings before forward to the encoder
+
+        return output
+
+    def set_imgsz(self, imgsz: list[int] = [1008, 1008]):
+        """Set the image size for the vision backbone."""
+        self.vision_backbone.set_imgsz(imgsz)

ultralytics/models/yolo/classify/val.py

@@ -57,7 +57,7 @@ class ClassificationValidator(BaseValidator):
         """Initialize ClassificationValidator with dataloader, save directory, and other parameters.
 
         Args:
-            dataloader (torch.utils.data.DataLoader, optional): Dataloader to use for validation.
+            dataloader (torch.utils.data.DataLoader, optional): DataLoader to use for validation.
             save_dir (str | Path, optional): Directory to save results.
             args (dict, optional): Arguments containing model and validation configuration.
             _callbacks (list, optional): List of callback functions to be called during validation.

ultralytics/models/yolo/detect/train.py

@@ -53,7 +53,7 @@ class DetectionTrainer(BaseTrainer):
     """
 
     def __init__(self, cfg=DEFAULT_CFG, overrides: dict[str, Any] | None = None, _callbacks=None):
-        """Initialize a DetectionTrainer object for training YOLO object detection model training.
+        """Initialize a DetectionTrainer object for training YOLO object detection models.
 
         Args:
             cfg (dict, optional): Default configuration dictionary containing training parameters.

ultralytics/models/yolo/detect/val.py

@@ -46,7 +46,7 @@ class DetectionValidator(BaseValidator):
         """Initialize detection validator with necessary variables and settings.
 
         Args:
-            dataloader (torch.utils.data.DataLoader, optional): Dataloader to use for validation.
+            dataloader (torch.utils.data.DataLoader, optional): DataLoader to use for validation.
             save_dir (Path, optional): Directory to save results.
             args (dict[str, Any], optional): Arguments for the validator.
             _callbacks (list[Any], optional): List of callback functions.
@@ -256,7 +256,7 @@ class DetectionValidator(BaseValidator):
         pf = "%22s" + "%11i" * 2 + "%11.3g" * len(self.metrics.keys)  # print format
         LOGGER.info(pf % ("all", self.seen, self.metrics.nt_per_class.sum(), *self.metrics.mean_results()))
         if self.metrics.nt_per_class.sum() == 0:
-            LOGGER.warning(f"no labels found in {self.args.task} set, can not compute metrics without labels")
+            LOGGER.warning(f"no labels found in {self.args.task} set, cannot compute metrics without labels")
 
         # Print results per class
         if self.args.verbose and not self.training and self.nc > 1 and len(self.metrics.stats):
@@ -308,7 +308,7 @@ class DetectionValidator(BaseValidator):
             batch_size (int): Size of each batch.
 
         Returns:
-            (torch.utils.data.DataLoader): Dataloader for validation.
+            (torch.utils.data.DataLoader): DataLoader for validation.
         """
         dataset = self.build_dataset(dataset_path, batch=batch_size, mode="val")
         return build_dataloader(
@@ -460,11 +460,11 @@ class DetectionValidator(BaseValidator):
 
         Args:
             stats (dict[str, Any]): Dictionary to store computed metrics and statistics.
-            pred_json (str | Path]): Path to JSON file containing predictions in COCO format.
-            anno_json (str | Path]): Path to JSON file containing ground truth annotations in COCO format.
-            iou_types (str | list[str]]): IoU type(s) for evaluation. Can be single string or list of strings. Common
+            pred_json (str | Path): Path to JSON file containing predictions in COCO format.
+            anno_json (str | Path): Path to JSON file containing ground truth annotations in COCO format.
+            iou_types (str | list[str]): IoU type(s) for evaluation. Can be single string or list of strings. Common
                 values include "bbox", "segm", "keypoints". Defaults to "bbox".
-            suffix (str | list[str]]): Suffix to append to metric names in stats dictionary. Should correspond to
+            suffix (str | list[str]): Suffix to append to metric names in stats dictionary. Should correspond to
                 iou_types if multiple types provided. Defaults to "Box".
 
         Returns:

ultralytics/models/yolo/obb/val.py

@@ -12,6 +12,7 @@ from ultralytics.models.yolo.detect import DetectionValidator
 from ultralytics.utils import LOGGER, ops
 from ultralytics.utils.metrics import OBBMetrics, batch_probiou
 from ultralytics.utils.nms import TorchNMS
+from ultralytics.utils.plotting import plot_images
 
 
 class OBBValidator(DetectionValidator):
@@ -49,7 +50,7 @@ class OBBValidator(DetectionValidator):
         extends the DetectionValidator class and configures it specifically for the OBB task.
 
         Args:
-            dataloader (torch.utils.data.DataLoader, optional): Dataloader to be used for validation.
+            dataloader (torch.utils.data.DataLoader, optional): DataLoader to be used for validation.
             save_dir (str | Path, optional): Directory to save results.
             args (dict | SimpleNamespace, optional): Arguments containing validation parameters.
             _callbacks (list, optional): List of callback functions to be called during validation.
@@ -141,24 +142,34 @@ class OBBValidator(DetectionValidator):
             "im_file": batch["im_file"][si],
         }
 
-    def plot_predictions(self, batch: dict[str, Any], preds: list[torch.Tensor], ni: int) -> None:
+    def plot_predictions(self, batch: dict[str, Any], preds: list[dict[str, torch.Tensor]], ni: int) -> None:
         """Plot predicted bounding boxes on input images and save the result.
 
         Args:
             batch (dict[str, Any]): Batch data containing images, file paths, and other metadata.
-            preds (list[torch.Tensor]): List of prediction tensors for each image in the batch.
+            preds (list[dict[str, torch.Tensor]]): List of prediction dictionaries for each image in the batch.
             ni (int): Batch index used for naming the output file.
 
         Examples:
             >>> validator = OBBValidator()
             >>> batch = {"img": images, "im_file": paths}
-            >>> preds = [torch.rand(10, 7)]  # Example predictions for one image
+            >>> preds = [{"bboxes": torch.rand(10, 5), "cls": torch.zeros(10), "conf": torch.rand(10)}]
             >>> validator.plot_predictions(batch, preds, 0)
         """
-        for p in preds:
-            # TODO: fix this duplicated `xywh2xyxy`
-            p["bboxes"][:, :4] = ops.xywh2xyxy(p["bboxes"][:, :4])  # convert to xyxy format for plotting
-        super().plot_predictions(batch, preds, ni)  # plot bboxes
+        if not preds:
+            return
+        for i, pred in enumerate(preds):
+            pred["batch_idx"] = torch.ones_like(pred["conf"]) * i
+        keys = preds[0].keys()
+        batched_preds = {k: torch.cat([x[k] for x in preds], dim=0) for k in keys}
+        plot_images(
+            images=batch["img"],
+            labels=batched_preds,
+            paths=batch["im_file"],
+            fname=self.save_dir / f"val_batch{ni}_pred.jpg",
+            names=self.names,
+            on_plot=self.on_plot,
+        )
 
     def pred_to_json(self, predn: dict[str, torch.Tensor], pbatch: dict[str, Any]) -> None:
         """Convert YOLO predictions to COCO JSON format with rotated bounding box information.

ultralytics/models/yolo/pose/val.py

@@ -59,7 +59,7 @@ class PoseValidator(DetectionValidator):
         specialized metrics for pose evaluation.
 
         Args:
-            dataloader (torch.utils.data.DataLoader, optional): Dataloader to be used for validation.
+            dataloader (torch.utils.data.DataLoader, optional): DataLoader to be used for validation.
             save_dir (Path | str, optional): Directory to save results.
             args (dict, optional): Arguments for the validator including task set to "pose".
             _callbacks (list, optional): List of callback functions to be executed during validation.

ultralytics/models/yolo/segment/val.py

@@ -39,7 +39,7 @@ class SegmentationValidator(DetectionValidator):
         """Initialize SegmentationValidator and set task to 'segment', metrics to SegmentMetrics.
 
         Args:
-            dataloader (torch.utils.data.DataLoader, optional): Dataloader to use for validation.
+            dataloader (torch.utils.data.DataLoader, optional): DataLoader to use for validation.
             save_dir (Path, optional): Directory to save results.
             args (namespace, optional): Arguments for the validator.
             _callbacks (list, optional): List of callback functions.

ultralytics/nn/autobackend.py

@@ -127,7 +127,7 @@ class AutoBackend(nn.Module):
 
     Methods:
         forward: Run inference on an input image.
-        from_numpy: Convert numpy array to tensor.
+        from_numpy: Convert NumPy arrays to tensors on the model device.
         warmup: Warm up the model with a dummy input.
         _model_type: Determine the model type from file path.
 
@@ -182,7 +182,7 @@ class AutoBackend(nn.Module):
             triton,
         ) = self._model_type("" if nn_module else model)
         fp16 &= pt or jit or onnx or xml or engine or nn_module or triton  # FP16
-        nhwc = coreml or saved_model or pb or tflite or edgetpu or rknn  # BHWC formats (vs torch BCWH)
+        nhwc = coreml or saved_model or pb or tflite or edgetpu or rknn  # BHWC formats (vs torch BCHW)
         stride, ch = 32, 3  # default stride and channels
         end2end, dynamic = False, False
         metadata, task = None, None
@@ -894,14 +894,14 @@ class AutoBackend(nn.Module):
         else:
             return self.from_numpy(y)
 
-    def from_numpy(self, x: np.ndarray) -> torch.Tensor:
-        """Convert a numpy array to a tensor.
+    def from_numpy(self, x: np.ndarray | torch.Tensor) -> torch.Tensor:
+        """Convert a NumPy array to a torch tensor on the model device.
 
         Args:
-            x (np.ndarray): The array to be converted.
+            x (np.ndarray | torch.Tensor): Input array or tensor.
 
         Returns:
-            (torch.Tensor): The converted tensor
+            (torch.Tensor): Tensor on `self.device`.
         """
         return torch.tensor(x).to(self.device) if isinstance(x, np.ndarray) else x
 
@@ -909,7 +909,7 @@ class AutoBackend(nn.Module):
         """Warm up the model by running one forward pass with a dummy input.
 
         Args:
-            imgsz (tuple): The shape of the dummy input tensor in the format (batch_size, channels, height, width)
+            imgsz (tuple[int, int, int, int]): Dummy input shape in (batch, channels, height, width) format.
         """
         warmup_types = self.pt, self.jit, self.onnx, self.engine, self.saved_model, self.pb, self.triton, self.nn_module
         if any(warmup_types) and (self.device.type != "cpu" or self.triton):
@@ -931,8 +931,8 @@ class AutoBackend(nn.Module):
             (list[bool]): List of booleans indicating the model type.
 
         Examples:
-            >>> model = AutoBackend(model="path/to/model.onnx")
-            >>> model_type = model._model_type()  # returns "onnx"
+            >>> types = AutoBackend._model_type("path/to/model.onnx")
+            >>> assert types[2]  # onnx
         """
         from ultralytics.engine.exporter import export_formats
 

ultralytics/nn/modules/block.py

@@ -1812,7 +1812,7 @@ class A2C2f(nn.Module):
         """
         super().__init__()
         c_ = int(c2 * e)  # hidden channels
-        assert c_ % 32 == 0, "Dimension of ABlock be a multiple of 32."
+        assert c_ % 32 == 0, "Dimension of ABlock must be a multiple of 32."
 
         self.cv1 = Conv(c1, c_, 1, 1)
         self.cv2 = Conv((1 + n) * c_, c2, 1)

ultralytics/nn/modules/transformer.py

@@ -359,7 +359,15 @@ class MLP(nn.Module):
     """
 
     def __init__(
-        self, input_dim: int, hidden_dim: int, output_dim: int, num_layers: int, act=nn.ReLU, sigmoid: bool = False
+        self,
+        input_dim: int,
+        hidden_dim: int,
+        output_dim: int,
+        num_layers: int,
+        act=nn.ReLU,
+        sigmoid: bool = False,
+        residual: bool = False,
+        out_norm: nn.Module = None,
     ):
         """Initialize the MLP with specified input, hidden, output dimensions and number of layers.
 
@@ -370,6 +378,8 @@ class MLP(nn.Module):
             num_layers (int): Number of layers.
             act (nn.Module): Activation function.
             sigmoid (bool): Whether to apply sigmoid to the output.
+            residual (bool): Whether to use residual connections.
+            out_norm (nn.Module, optional): Normalization layer for the output.
         """
         super().__init__()
         self.num_layers = num_layers
@@ -377,6 +387,12 @@ class MLP(nn.Module):
         self.layers = nn.ModuleList(nn.Linear(n, k) for n, k in zip([input_dim, *h], [*h, output_dim]))
         self.sigmoid = sigmoid
         self.act = act()
+        if residual and input_dim != output_dim:
+            raise ValueError("residual is only supported if input_dim == output_dim")
+        self.residual = residual
+        # whether to apply a normalization layer to the output
+        assert isinstance(out_norm, nn.Module) or out_norm is None
+        self.out_norm = out_norm or nn.Identity()
 
     def forward(self, x: torch.Tensor) -> torch.Tensor:
         """Forward pass for the entire MLP.
@@ -387,8 +403,12 @@ class MLP(nn.Module):
         Returns:
             (torch.Tensor): Output tensor after MLP.
         """
+        orig_x = x
         for i, layer in enumerate(self.layers):
             x = getattr(self, "act", nn.ReLU())(layer(x)) if i < self.num_layers - 1 else layer(x)
+        if getattr(self, "residual", False):
+            x = x + orig_x
+        x = getattr(self, "out_norm", nn.Identity())(x)
         return x.sigmoid() if getattr(self, "sigmoid", False) else x
 
 

ultralytics/nn/tasks.py

@@ -866,7 +866,7 @@ class WorldModel(DetectionModel):
         self.model[-1].nc = len(text)
 
     def get_text_pe(self, text, batch=80, cache_clip_model=True):
-        """Set classes in advance so that model could do offline-inference without clip model.
+        """Get text positional embeddings for offline inference without CLIP model.
 
         Args:
             text (list[str]): List of class names.
@@ -987,13 +987,13 @@ class YOLOEModel(DetectionModel):
 
     @smart_inference_mode()
     def get_text_pe(self, text, batch=80, cache_clip_model=False, without_reprta=False):
-        """Set classes in advance so that model could do offline-inference without clip model.
+        """Get text positional embeddings for offline inference without CLIP model.
 
         Args:
             text (list[str]): List of class names.
             batch (int): Batch size for processing text tokens.
             cache_clip_model (bool): Whether to cache the CLIP model.
-            without_reprta (bool): Whether to return text embeddings cooperated with reprta module.
+            without_reprta (bool): Whether to return text embeddings without reprta module processing.
 
         Returns:
             (torch.Tensor): Text positional embeddings.

ultralytics/nn/text_model.py

@@ -196,12 +196,7 @@ class MobileCLIP(TextModel):
             device (torch.device): Device to load the model on.
         """
         try:
-            import warnings
-
-            # Suppress 'timm.models.layers is deprecated, please import via timm.layers' warning from mobileclip usage
-            with warnings.catch_warnings():
-                warnings.filterwarnings("ignore", category=FutureWarning)
-                import mobileclip
+            import mobileclip
         except ImportError:
             # Ultralytics fork preferred since Apple MobileCLIP repo has incorrect version of torchvision
             checks.check_requirements("git+https://github.com/ultralytics/mobileclip.git")
@@ -308,7 +303,7 @@ class MobileCLIPTS(TextModel):
             (torch.Tensor): Tokenized text inputs with shape (batch_size, sequence_length).
 
         Examples:
-            >>> model = MobileCLIPTS("cpu")
+            >>> model = MobileCLIPTS(device=torch.device("cpu"))
             >>> tokens = model.tokenize(["a photo of a cat", "a photo of a dog"])
             >>> strict_tokens = model.tokenize(
             ...     ["a very long caption"], truncate=False

ultralytics/solutions/ai_gym.py

@@ -13,7 +13,7 @@ class AIGym(BaseSolution):
     repetitions of exercises based on predefined angle thresholds for up and down positions.
 
     Attributes:
-        states (dict[float, int, str]): Stores per-track angle, count, and stage for workout monitoring.
+        states (dict[int, dict[str, float | int | str]]): Per-track angle, rep count, and stage for workout monitoring.
         up_angle (float): Angle threshold for considering the 'up' position of an exercise.
         down_angle (float): Angle threshold for considering the 'down' position of an exercise.
         kpts (list[int]): Indices of keypoints used for angle calculation.

ultralytics/solutions/analytics.py

@@ -56,7 +56,7 @@ class Analytics(BaseSolution):
         from matplotlib.backends.backend_agg import FigureCanvasAgg
         from matplotlib.figure import Figure
 
-        self.type = self.CFG["analytics_type"]  # type of analytics i.e "line", "pie", "bar" or "area" charts.
+        self.type = self.CFG["analytics_type"]  # Chart type: "line", "pie", "bar", or "area".
         self.x_label = "Classes" if self.type in {"bar", "pie"} else "Frame#"
         self.y_label = "Total Counts"
 
@@ -66,10 +66,10 @@ class Analytics(BaseSolution):
         self.title = "Ultralytics Solutions"  # window name
         self.max_points = 45  # maximum points to be drawn on window
         self.fontsize = 25  # text font size for display
-        figsize = self.CFG["figsize"]  # set output image size i.e (12.8, 7.2) -> w = 1280, h = 720
+        figsize = self.CFG["figsize"]  # Output size, e.g. (12.8, 7.2) -> 1280x720.
         self.color_cycle = cycle(["#DD00BA", "#042AFF", "#FF4447", "#7D24FF", "#BD00FF"])
 
-        self.total_counts = 0  # count variable for storing total counts i.e. for line
+        self.total_counts = 0  # Stores total counts for line charts.
         self.clswise_count = {}  # dictionary for class-wise counts
         self.update_every = kwargs.get("update_every", 30)  # Only update graph every 30 frames by default
         self.last_plot_im = None  # Cache of the last rendered chart
@@ -104,7 +104,7 @@ class Analytics(BaseSolution):
                 and 'classwise_count' (dict, per-class object count).
 
         Raises:
-            ModuleNotFoundError: If an unsupported chart type is specified.
+            ValueError: If an unsupported chart type is specified.
 
         Examples:
             >>> analytics = Analytics(analytics_type="line")
@@ -131,9 +131,9 @@ class Analytics(BaseSolution):
                 )
             plot_im = self.last_plot_im
         else:
-            raise ModuleNotFoundError(f"{self.type} chart is not supported ❌")
+            raise ValueError(f"Unsupported analytics_type='{self.type}'. Supported types: line, bar, pie, area.")
 
-        # return output dictionary with summary for more usage
+        # Return results for downstream use.
         return SolutionResults(plot_im=plot_im, total_tracks=len(self.track_ids), classwise_count=self.clswise_count)
 
     def update_graph(

ultralytics/solutions/config.py

@@ -35,7 +35,7 @@ class SolutionConfig:
         vision_point (tuple[int, int]): Reference point for directional tracking or perspective drawing.
         crop_dir (str): Directory path to save cropped detection images.
         json_file (str): Path to a JSON file containing data for parking areas.
-        line_width (int): Width for visual display i.e. bounding boxes, keypoints, counts.
+        line_width (int): Width for visual display, e.g. bounding boxes, keypoints, and counts.
         records (int): Number of detection records to send email alerts.
         fps (float): Frame rate (Frames Per Second) for speed estimation calculation.
         max_hist (int): Maximum number of historical points or states stored per tracked object for speed estimation.

ultralytics/solutions/distance_calculation.py

@@ -17,7 +17,7 @@ class DistanceCalculation(BaseSolution):
 
     Attributes:
         left_mouse_count (int): Counter for left mouse button clicks.
-        selected_boxes (dict[int, list[float]]): Dictionary to store selected bounding boxes and their track IDs.
+        selected_boxes (dict[int, Any]): Dictionary to store selected bounding boxes keyed by track ID.
         centroids (list[list[int]]): List to store centroids of selected bounding boxes.
 
     Methods:

ultralytics/solutions/object_counter.py

@@ -19,7 +19,7 @@ class ObjectCounter(BaseSolution):
         in_count (int): Counter for objects moving inward.
         out_count (int): Counter for objects moving outward.
         counted_ids (list[int]): List of IDs of objects that have been counted.
-        classwise_counts (dict[str, dict[str, int]]): Dictionary for counts, categorized by object class.
+        classwise_count (dict[str, dict[str, int]]): Dictionary for counts, categorized by object class.
         region_initialized (bool): Flag indicating whether the counting region has been initialized.
         show_in (bool): Flag to control display of inward count.
         show_out (bool): Flag to control display of outward count.

ultralytics/solutions/object_cropper.py

@@ -1,6 +1,5 @@
 # Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license
 
-import os
 from pathlib import Path
 from typing import Any
 
@@ -40,12 +39,10 @@ class ObjectCropper(BaseSolution):
         super().__init__(**kwargs)
 
         self.crop_dir = self.CFG["crop_dir"]  # Directory for storing cropped detections
-        if not os.path.exists(self.crop_dir):
-            os.mkdir(self.crop_dir)  # Create directory if it does not exist
+        Path(self.crop_dir).mkdir(parents=True, exist_ok=True)
         if self.CFG["show"]:
-            self.LOGGER.warning(
-                f"show=True disabled for crop solution, results will be saved in the directory named: {self.crop_dir}"
-            )
+            self.LOGGER.warning(f"show=True is not supported for ObjectCropper; saving crops to '{self.crop_dir}'.")
+            self.CFG["show"] = False
         self.crop_idx = 0  # Initialize counter for total cropped objects
         self.iou = self.CFG["iou"]
         self.conf = self.CFG["conf"]