ultralytics 8.1.15__tar.gz → 8.1.17__tar.gz

{ultralytics-8.1.15/ultralytics.egg-info → ultralytics-8.1.17}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: ultralytics
-Version: 8.1.15
+Version: 8.1.17
 Summary: Ultralytics YOLOv8 for SOTA object detection, multi-object tracking, instance segmentation, pose estimation and image classification.
 Author: Glenn Jocher, Ayush Chaurasia, Jing Qiu
 Maintainer: Glenn Jocher, Ayush Chaurasia, Jing Qiu
@@ -55,7 +55,7 @@ Requires-Dist: mkdocs-material>=9.5.9; extra == "dev"
 Requires-Dist: mkdocstrings[python]; extra == "dev"
 Requires-Dist: mkdocs-jupyter; extra == "dev"
 Requires-Dist: mkdocs-redirects; extra == "dev"
-Requires-Dist: mkdocs-ultralytics-plugin>=0.0.43; extra == "dev"
+Requires-Dist: mkdocs-ultralytics-plugin>=0.0.44; extra == "dev"
 Provides-Extra: export
 Requires-Dist: onnx>=1.12.0; extra == "export"
 Requires-Dist: coremltools>=7.0; (platform_system != "Windows" and python_version <= "3.11") and extra == "export"

{ultralytics-8.1.15 → ultralytics-8.1.17}/pyproject.toml

@@ -93,7 +93,7 @@ dev = [
     "mkdocstrings[python]",
     "mkdocs-jupyter", # for notebooks
     "mkdocs-redirects", # for 301 redirects
-    "mkdocs-ultralytics-plugin>=0.0.43", # for meta descriptions and images, dates and authors
+    "mkdocs-ultralytics-plugin>=0.0.44", # for meta descriptions and images, dates and authors
 ]
 export = [
     "onnx>=1.12.0", # ONNX export

{ultralytics-8.1.15 → ultralytics-8.1.17}/tests/test_explorer.py

@@ -1,11 +1,13 @@
 # Ultralytics YOLO 🚀, AGPL-3.0 license
 
 import PIL
+import pytest
 
 from ultralytics import Explorer
 from ultralytics.utils import ASSETS
 
 
+@pytest.mark.slow
 def test_similarity():
     """Test similarity calculations and SQL queries for correctness and response length."""
     exp = Explorer()
@@ -22,6 +24,7 @@ def test_similarity():
     assert len(sql) > 0
 
 
+@pytest.mark.slow
 def test_det():
     """Test detection functionalities and ensure the embedding table has bounding boxes."""
     exp = Explorer(data="coco8.yaml", model="yolov8n.pt")
@@ -34,6 +37,7 @@ def test_det():
     assert isinstance(similar, PIL.Image.Image)
 
 
+@pytest.mark.slow
 def test_seg():
     """Test segmentation functionalities and verify the embedding table includes masks."""
     exp = Explorer(data="coco8-seg.yaml", model="yolov8n-seg.pt")
@@ -45,6 +49,7 @@ def test_seg():
     assert isinstance(similar, PIL.Image.Image)
 
 
+@pytest.mark.slow
 def test_pose():
     """Test pose estimation functionalities and check the embedding table for keypoints."""
     exp = Explorer(data="coco8-pose.yaml", model="yolov8n-pose.pt")

{ultralytics-8.1.15 → ultralytics-8.1.17}/ultralytics/__init__.py

@@ -1,6 +1,6 @@
 # Ultralytics YOLO 🚀, AGPL-3.0 license
 
-__version__ = "8.1.15"
+__version__ = "8.1.17"
 
 from ultralytics.data.explorer.explorer import Explorer
 from ultralytics.models import RTDETR, SAM, YOLO, YOLOWorld

{ultralytics-8.1.15 → ultralytics-8.1.17}/ultralytics/cfg/default.yaml

@@ -9,7 +9,7 @@ model: # (str, optional) path to model file, i.e. yolov8n.pt, yolov8n.yaml
 data: # (str, optional) path to data file, i.e. coco128.yaml
 epochs: 100 # (int) number of epochs to train for
 time: # (float, optional) number of hours to train for, overrides epochs if supplied
-patience: 50 # (int) epochs to wait for no observable improvement for early stopping of training
+patience: 100 # (int) epochs to wait for no observable improvement for early stopping of training
 batch: 16 # (int) number of images per batch (-1 for AutoBatch)
 imgsz: 640 # (int | list) input images size as int for train and val modes, or list[w,h] for predict and export modes
 save: True # (bool) save train checkpoints and predict results

{ultralytics-8.1.15 → ultralytics-8.1.17}/ultralytics/data/converter.py

@@ -418,8 +418,8 @@ def min_index(arr1, arr2):
     Find a pair of indexes with the shortest distance between two arrays of 2D points.
 
     Args:
-        arr1 (np.array): A NumPy array of shape (N, 2) representing N 2D points.
-        arr2 (np.array): A NumPy array of shape (M, 2) representing M 2D points.
+        arr1 (np.ndarray): A NumPy array of shape (N, 2) representing N 2D points.
+        arr2 (np.ndarray): A NumPy array of shape (M, 2) representing M 2D points.
 
     Returns:
         (tuple): A tuple containing the indexes of the points with the shortest distance in arr1 and arr2 respectively.

{ultralytics-8.1.15 → ultralytics-8.1.17}/ultralytics/data/dataset.py

@@ -226,35 +226,42 @@ class YOLODataset(BaseDataset):
 # Classification dataloaders -------------------------------------------------------------------------------------------
 class ClassificationDataset(torchvision.datasets.ImageFolder):
     """
-    YOLO Classification Dataset.
+    Extends torchvision ImageFolder to support YOLO classification tasks, offering functionalities like image
+    augmentation, caching, and verification. It's designed to efficiently handle large datasets for training deep
+    learning models, with optional image transformations and caching mechanisms to speed up training.
 
-    Args:
-        root (str): Dataset path.
+    This class allows for augmentations using both torchvision and Albumentations libraries, and supports caching images
+    in RAM or on disk to reduce IO overhead during training. Additionally, it implements a robust verification process
+    to ensure data integrity and consistency.
 
     Attributes:
-        cache_ram (bool): True if images should be cached in RAM, False otherwise.
-        cache_disk (bool): True if images should be cached on disk, False otherwise.
-        samples (list): List of samples containing file, index, npy, and im.
-        torch_transforms (callable): torchvision transforms applied to the dataset.
-        album_transforms (callable, optional): Albumentations transforms applied to the dataset if augment is True.
+        cache_ram (bool): Indicates if caching in RAM is enabled.
+        cache_disk (bool): Indicates if caching on disk is enabled.
+        samples (list): A list of tuples, each containing the path to an image, its class index, path to its .npy cache
+                        file (if caching on disk), and optionally the loaded image array (if caching in RAM).
+        torch_transforms (callable): PyTorch transforms to be applied to the images.
     """
 
-    def __init__(self, root, args, augment=False, cache=False, prefix=""):
+    def __init__(self, root, args, augment=False, prefix=""):
         """
         Initialize YOLO object with root, image size, augmentations, and cache settings.
 
         Args:
-            root (str): Dataset path.
-            args (Namespace): Argument parser containing dataset related settings.
-            augment (bool, optional): True if dataset should be augmented, False otherwise. Defaults to False.
-            cache (bool | str | optional): Cache setting, can be True, False, 'ram' or 'disk'. Defaults to False.
+            root (str): Path to the dataset directory where images are stored in a class-specific folder structure.
+            args (Namespace): Configuration containing dataset-related settings such as image size, augmentation
+                parameters, and cache settings. It includes attributes like `imgsz` (image size), `fraction` (fraction
+                of data to use), `scale`, `fliplr`, `flipud`, `cache` (disk or RAM caching for faster training),
+                `auto_augment`, `hsv_h`, `hsv_s`, `hsv_v`, and `crop_fraction`.
+            augment (bool, optional): Whether to apply augmentations to the dataset. Default is False.
+            prefix (str, optional): Prefix for logging and cache filenames, aiding in dataset identification and
+                debugging. Default is an empty string.
         """
         super().__init__(root=root)
         if augment and args.fraction < 1.0:  # reduce training fraction
             self.samples = self.samples[: round(len(self.samples) * args.fraction)]
         self.prefix = colorstr(f"{prefix}: ") if prefix else ""
-        self.cache_ram = cache is True or cache == "ram"
-        self.cache_disk = cache == "disk"
+        self.cache_ram = args.cache is True or args.cache == "ram"  # cache images into RAM
+        self.cache_disk = args.cache == "disk"  # cache images on hard drive as uncompressed *.npy files
         self.samples = self.verify_images()  # filter out bad images
         self.samples = [list(x) + [Path(x[0]).with_suffix(".npy"), None] for x in self.samples]  # file, index, npy, im
         scale = (1.0 - args.scale, 1.0)  # (0.08, 1.0)

{ultralytics-8.1.15 → ultralytics-8.1.17}/ultralytics/data/utils.py

@@ -467,7 +467,6 @@ class HUBDatasetStats:
 
         self.hub_dir = Path(f'{data["path"]}-hub')
         self.im_dir = self.hub_dir / "images"
-        self.im_dir.mkdir(parents=True, exist_ok=True)  # makes /images
         self.stats = {"nc": len(data["names"]), "names": list(data["names"].values())}  # statistics dictionary
         self.data = data
 
@@ -551,6 +550,7 @@ class HUBDatasetStats:
 
         # Save, print and return
         if save:
+            self.hub_dir.mkdir(parents=True, exist_ok=True)  # makes dataset-hub/
             stats_path = self.hub_dir / "stats.json"
             LOGGER.info(f"Saving {stats_path.resolve()}...")
             with open(stats_path, "w") as f:
@@ -563,6 +563,7 @@ class HUBDatasetStats:
         """Compress images for Ultralytics HUB."""
         from ultralytics.data import YOLODataset  # ClassificationDataset
 
+        self.im_dir.mkdir(parents=True, exist_ok=True)  # makes dataset-hub/images/
         for split in "train", "val", "test":
             if self.data.get(split) is None:
                 continue

{ultralytics-8.1.15 → ultralytics-8.1.17}/ultralytics/engine/exporter.py

@@ -716,12 +716,13 @@ class Exporter:
             import tensorflow as tf  # noqa
         check_requirements(
             (
-                "onnx",
+                "onnx>=1.12.0",
                 "onnx2tf>=1.15.4,<=1.17.5",
                 "sng4onnx>=1.0.1",
                 "onnxsim>=0.4.33",
                 "onnx_graphsurgeon>=0.3.26",
                 "tflite_support",
+                "flatbuffers>=23.5.26",  # update old 'flatbuffers' included inside tensorflow package
                 "onnxruntime-gpu" if cuda else "onnxruntime",
             ),
             cmds="--extra-index-url https://pypi.ngc.nvidia.com",
@@ -860,8 +861,7 @@ class Exporter:
     @try_export
     def export_tfjs(self, prefix=colorstr("TensorFlow.js:")):
         """YOLOv8 TensorFlow.js export."""
-        # JAX bug requiring install constraints in https://github.com/google/jax/issues/18978
-        check_requirements(["jax<=0.4.21", "jaxlib<=0.4.21", "tensorflowjs"])
+        check_requirements("tensorflowjs")
         import tensorflow as tf
         import tensorflowjs as tfjs  # noqa
 

{ultralytics-8.1.15 → ultralytics-8.1.17}/ultralytics/engine/model.py

@@ -5,6 +5,9 @@ import sys
 from pathlib import Path
 from typing import Union
 
+import numpy as np
+import torch
+
 from ultralytics.cfg import TASK2DATA, get_cfg, get_save_dir
 from ultralytics.hub.utils import HUB_WEB_ROOT
 from ultralytics.nn.tasks import attempt_load_one_weight, guess_model_task, nn, yaml_model_load
@@ -78,7 +81,12 @@ class Model(nn.Module):
         NotImplementedError: If a specific model task or mode is not supported.
     """
 
-    def __init__(self, model: Union[str, Path] = "yolov8n.pt", task=None, verbose=False) -> None:
+    def __init__(
+        self,
+        model: Union[str, Path] = "yolov8n.pt",
+        task: str = None,
+        verbose: bool = False,
+    ) -> None:
         """
         Initializes a new instance of the YOLO model class.
 
@@ -135,7 +143,12 @@ class Model(nn.Module):
 
         self.model_name = model
 
-    def __call__(self, source=None, stream=False, **kwargs):
+    def __call__(
+        self,
+        source: Union[str, Path, int, list, tuple, np.ndarray, torch.Tensor] = None,
+        stream: bool = False,
+        **kwargs,
+    ) -> list:
         """
         An alias for the predict method, enabling the model instance to be callable.
 
@@ -143,8 +156,9 @@ class Model(nn.Module):
         with the required arguments for prediction.
 
         Args:
-            source (str | int | PIL.Image | np.ndarray, optional): The source of the image for making predictions.
-                Accepts various types, including file paths, URLs, PIL images, and numpy arrays. Defaults to None.
+            source (str | Path | int | PIL.Image | np.ndarray, optional): The source of the image for making
+                predictions. Accepts various types, including file paths, URLs, PIL images, and numpy arrays.
+                Defaults to None.
             stream (bool, optional): If True, treats the input source as a continuous stream for predictions.
                 Defaults to False.
             **kwargs (dict): Additional keyword arguments for configuring the prediction process.
@@ -163,7 +177,7 @@ class Model(nn.Module):
         return session if session.client.authenticated else None
 
     @staticmethod
-    def is_triton_model(model):
+    def is_triton_model(model: str) -> bool:
         """Is model a Triton Server URL string, i.e. <scheme>://<netloc>/<endpoint>/<task_name>"""
         from urllib.parse import urlsplit
 
@@ -171,7 +185,7 @@ class Model(nn.Module):
         return url.netloc and url.path and url.scheme in {"http", "grpc"}
 
     @staticmethod
-    def is_hub_model(model):
+    def is_hub_model(model: str) -> bool:
         """Check if the provided model is a HUB model."""
         return any(
             (
@@ -181,7 +195,7 @@ class Model(nn.Module):
             )
         )
 
-    def _new(self, cfg: str, task=None, model=None, verbose=False):
+    def _new(self, cfg: str, task=None, model=None, verbose=False) -> None:
         """
         Initializes a new model and infers the task type from the model definitions.
 
@@ -202,7 +216,7 @@ class Model(nn.Module):
         self.model.args = {**DEFAULT_CFG_DICT, **self.overrides}  # combine default and model args (prefer model args)
         self.model.task = self.task
 
-    def _load(self, weights: str, task=None):
+    def _load(self, weights: str, task=None) -> None:
         """
         Initializes a new model and infers the task type from the model head.
 
@@ -224,7 +238,7 @@ class Model(nn.Module):
         self.overrides["model"] = weights
         self.overrides["task"] = self.task
 
-    def _check_is_pytorch_model(self):
+    def _check_is_pytorch_model(self) -> None:
         """Raises TypeError is model is not a PyTorch model."""
         pt_str = isinstance(self.model, (str, Path)) and Path(self.model).suffix == ".pt"
         pt_module = isinstance(self.model, nn.Module)
@@ -237,7 +251,7 @@ class Model(nn.Module):
                 f"argument directly in your inference command, i.e. 'model.predict(source=..., device=0)'"
             )
 
-    def reset_weights(self):
+    def reset_weights(self) -> "Model":
         """
         Resets the model parameters to randomly initialized values, effectively discarding all training information.
 
@@ -259,7 +273,7 @@ class Model(nn.Module):
             p.requires_grad = True
         return self
 
-    def load(self, weights="yolov8n.pt"):
+    def load(self, weights: Union[str, Path] = "yolov8n.pt") -> "Model":
         """
         Loads parameters from the specified weights file into the model.
 
@@ -281,24 +295,22 @@ class Model(nn.Module):
         self.model.load(weights)
         return self
 
-    def save(self, filename="model.pt"):
+    def save(self, filename: Union[str, Path] = "saved_model.pt") -> None:
         """
         Saves the current model state to a file.
 
         This method exports the model's checkpoint (ckpt) to the specified filename.
 
         Args:
-            filename (str): The name of the file to save the model to. Defaults to 'model.pt'.
+            filename (str | Path): The name of the file to save the model to. Defaults to 'saved_model.pt'.
 
         Raises:
             AssertionError: If the model is not a PyTorch model.
         """
         self._check_is_pytorch_model()
-        import torch
-
         torch.save(self.ckpt, filename)
 
-    def info(self, detailed=False, verbose=True):
+    def info(self, detailed: bool = False, verbose: bool = True):
         """
         Logs or returns model information.
 
@@ -330,7 +342,12 @@ class Model(nn.Module):
         self._check_is_pytorch_model()
         self.model.fuse()
 
-    def embed(self, source=None, stream=False, **kwargs):
+    def embed(
+        self,
+        source: Union[str, Path, int, list, tuple, np.ndarray, torch.Tensor] = None,
+        stream: bool = False,
+        **kwargs,
+    ) -> list:
         """
         Generates image embeddings based on the provided source.
 
@@ -353,7 +370,13 @@ class Model(nn.Module):
             kwargs["embed"] = [len(self.model.model) - 2]  # embed second-to-last layer if no indices passed
         return self.predict(source, stream, **kwargs)
 
-    def predict(self, source=None, stream=False, predictor=None, **kwargs):
+    def predict(
+        self,
+        source: Union[str, Path, int, list, tuple, np.ndarray, torch.Tensor] = None,
+        stream: bool = False,
+        predictor=None,
+        **kwargs,
+    ) -> list:
         """
         Performs predictions on the given image source using the YOLO model.
 
@@ -405,7 +428,13 @@ class Model(nn.Module):
             self.predictor.set_prompts(prompts)
         return self.predictor.predict_cli(source=source) if is_cli else self.predictor(source=source, stream=stream)
 
-    def track(self, source=None, stream=False, persist=False, **kwargs):
+    def track(
+        self,
+        source: Union[str, Path, int, list, tuple, np.ndarray, torch.Tensor] = None,
+        stream: bool = False,
+        persist: bool = False,
+        **kwargs,
+    ) -> list:
         """
         Conducts object tracking on the specified input source using the registered trackers.
 
@@ -438,7 +467,11 @@ class Model(nn.Module):
         kwargs["mode"] = "track"
         return self.predict(source=source, stream=stream, **kwargs)
 
-    def val(self, validator=None, **kwargs):
+    def val(
+        self,
+        validator=None,
+        **kwargs,
+    ):
         """
         Validates the model using a specified dataset and validation configuration.
 
@@ -471,7 +504,10 @@ class Model(nn.Module):
         self.metrics = validator.metrics
         return validator.metrics
 
-    def benchmark(self, **kwargs):
+    def benchmark(
+        self,
+        **kwargs,
+    ):
         """
         Benchmarks the model across various export formats to evaluate performance.
 
@@ -509,7 +545,10 @@ class Model(nn.Module):
             verbose=kwargs.get("verbose"),
         )
 
-    def export(self, **kwargs):
+    def export(
+        self,
+        **kwargs,
+    ):
         """
         Exports the model to a different format suitable for deployment.
 
@@ -537,7 +576,11 @@ class Model(nn.Module):
         args = {**self.overrides, **custom, **kwargs, "mode": "export"}  # highest priority args on the right
         return Exporter(overrides=args, _callbacks=self.callbacks)(model=self.model)
 
-    def train(self, trainer=None, **kwargs):
+    def train(
+        self,
+        trainer=None,
+        **kwargs,
+    ):
         """
         Trains the model using the specified dataset and training configuration.
 
@@ -607,7 +650,13 @@ class Model(nn.Module):
             self.metrics = getattr(self.trainer.validator, "metrics", None)  # TODO: no metrics returned by DDP
         return self.metrics
 
-    def tune(self, use_ray=False, iterations=10, *args, **kwargs):
+    def tune(
+        self,
+        use_ray=False,
+        iterations=10,
+        *args,
+        **kwargs,
+    ):
         """
         Conducts hyperparameter tuning for the model, with an option to use Ray Tune.
 
@@ -640,7 +689,7 @@ class Model(nn.Module):
             args = {**self.overrides, **custom, **kwargs, "mode": "train"}  # highest priority args on the right
             return Tuner(args=args, _callbacks=self.callbacks)(model=self, iterations=iterations)
 
-    def _apply(self, fn):
+    def _apply(self, fn) -> "Model":
         """Apply to(), cpu(), cuda(), half(), float() to model tensors that are not parameters or registered buffers."""
         self._check_is_pytorch_model()
         self = super()._apply(fn)  # noqa
@@ -649,7 +698,7 @@ class Model(nn.Module):
         return self
 
     @property
-    def names(self):
+    def names(self) -> list:
         """
         Retrieves the class names associated with the loaded model.
 
@@ -664,7 +713,7 @@ class Model(nn.Module):
         return check_class_names(self.model.names) if hasattr(self.model, "names") else None
 
     @property
-    def device(self):
+    def device(self) -> torch.device:
         """
         Retrieves the device on which the model's parameters are allocated.
 
@@ -688,7 +737,7 @@ class Model(nn.Module):
         """
         return self.model.transforms if hasattr(self.model, "transforms") else None
 
-    def add_callback(self, event: str, func):
+    def add_callback(self, event: str, func) -> None:
         """
         Adds a callback function for a specified event.
 
@@ -704,7 +753,7 @@ class Model(nn.Module):
         """
         self.callbacks[event].append(func)
 
-    def clear_callback(self, event: str):
+    def clear_callback(self, event: str) -> None:
         """
         Clears all callback functions registered for a specified event.
 
@@ -718,7 +767,7 @@ class Model(nn.Module):
         """
         self.callbacks[event] = []
 
-    def reset_callbacks(self):
+    def reset_callbacks(self) -> None:
         """
         Resets all callbacks to their default functions.
 
@@ -729,7 +778,7 @@ class Model(nn.Module):
             self.callbacks[event] = [callbacks.default_callbacks[event][0]]
 
     @staticmethod
-    def _reset_ckpt_args(args):
+    def _reset_ckpt_args(args: dict) -> dict:
         """Reset arguments when loading a PyTorch model."""
         include = {"imgsz", "data", "task", "single_cls"}  # only remember these arguments when loading a PyTorch model
         return {k: v for k, v in args.items() if k in include}
@@ -739,7 +788,7 @@ class Model(nn.Module):
     #    name = self.__class__.__name__
     #    raise AttributeError(f"'{name}' object has no attribute '{attr}'. See valid attributes below.\n{self.__doc__}")
 
-    def _smart_load(self, key):
+    def _smart_load(self, key: str):
         """Load model/trainer/validator/predictor."""
         try:
             return self.task_map[self.task][key]
@@ -751,7 +800,7 @@ class Model(nn.Module):
             ) from e
 
     @property
-    def task_map(self):
+    def task_map(self) -> dict:
         """
         Map head to model, trainer, validator, and predictor classes.
 

{ultralytics-8.1.15 → ultralytics-8.1.17}/ultralytics/engine/trainer.py

@@ -252,7 +252,7 @@ class BaseTrainer:
             if any(x in k for x in freeze_layer_names):
                 LOGGER.info(f"Freezing layer '{k}'")
                 v.requires_grad = False
-            elif not v.requires_grad:
+            elif not v.requires_grad and v.dtype.is_floating_point:  # only floating point Tensor can require gradients
                 LOGGER.info(
                     f"WARNING ⚠️ setting 'requires_grad=True' for frozen layer '{k}'. "
                     "See ultralytics.engine.trainer for customization of frozen layers."

{ultralytics-8.1.15 → ultralytics-8.1.17}/ultralytics/models/yolo/detect/val.py

@@ -132,8 +132,7 @@ class DetectionValidator(BaseValidator):
                 if nl:
                     for k in self.stats.keys():
                         self.stats[k].append(stat[k])
-                    # TODO: obb has not supported confusion_matrix yet.
-                    if self.args.plots and self.args.task != "obb":
+                    if self.args.plots:
                         self.confusion_matrix.process_batch(detections=None, gt_bboxes=bbox, gt_cls=cls)
                 continue
 
@@ -147,8 +146,7 @@ class DetectionValidator(BaseValidator):
             # Evaluate
             if nl:
                 stat["tp"] = self._process_batch(predn, bbox, cls)
-                # TODO: obb has not supported confusion_matrix yet.
-                if self.args.plots and self.args.task != "obb":
+                if self.args.plots:
                     self.confusion_matrix.process_batch(predn, bbox, cls)
             for k in self.stats.keys():
                 self.stats[k].append(stat[k])

{ultralytics-8.1.15 → ultralytics-8.1.17}/ultralytics/models/yolo/obb/val.py

@@ -55,10 +55,11 @@ class OBBValidator(DetectionValidator):
         Return correct prediction matrix.
 
         Args:
-            detections (torch.Tensor): Tensor of shape [N, 6] representing detections.
-                Each detection is of the format: x1, y1, x2, y2, conf, class.
-            labels (torch.Tensor): Tensor of shape [M, 5] representing labels.
-                Each label is of the format: class, x1, y1, x2, y2.
+            detections (torch.Tensor): Tensor of shape [N, 7] representing detections.
+                Each detection is of the format: x1, y1, x2, y2, conf, class, angle.
+            gt_bboxes (torch.Tensor): Tensor of shape [M, 5] representing rotated boxes.
+                Each box is of the format: x1, y1, x2, y2, angle.
+            labels (torch.Tensor): Tensor of shape [M] representing labels.
 
         Returns:
             (torch.Tensor): Correct prediction matrix of shape [N, 10] for 10 IoU levels.

{ultralytics-8.1.15 → ultralytics-8.1.17}/ultralytics/nn/tasks.py

@@ -761,6 +761,8 @@ def attempt_load_weights(weights, device=None, inplace=True, fuse=False):
     for m in ensemble.modules():
         if hasattr(m, "inplace"):
             m.inplace = inplace
+        elif isinstance(m, nn.Upsample) and not hasattr(m, "recompute_scale_factor"):
+            m.recompute_scale_factor = None  # torch 1.11.0 compatibility
 
     # Return model
     if len(ensemble) == 1:
@@ -794,6 +796,8 @@ def attempt_load_one_weight(weight, device=None, inplace=True, fuse=False):
     for m in model.modules():
         if hasattr(m, "inplace"):
             m.inplace = inplace
+        elif isinstance(m, nn.Upsample) and not hasattr(m, "recompute_scale_factor"):
+            m.recompute_scale_factor = None  # torch 1.11.0 compatibility
 
     # Return model and ckpt
     return model, ckpt

{ultralytics-8.1.15 → ultralytics-8.1.17}/ultralytics/solutions/speed_estimation.py

@@ -154,13 +154,14 @@ class SpeedEstimator:
         self.trk_previous_times[trk_id] = time()
         self.trk_previous_points[trk_id] = track[-1]
 
-    def estimate_speed(self, im0, tracks):
+    def estimate_speed(self, im0, tracks, region_color=(255, 0, 0)):
         """
         Calculate object based on tracking data.
 
         Args:
             im0 (nd array): Image
             tracks (list): List of tracks obtained from the object tracking process.
+            region_color (tuple): Color to use when drawing regions.
         """
         self.im0 = im0
         if tracks[0].boxes.id is None:
@@ -170,7 +171,7 @@ class SpeedEstimator:
         self.extract_tracks(tracks)
 
         self.annotator = Annotator(self.im0, line_width=2)
-        self.annotator.draw_region(reg_pts=self.reg_pts, color=(255, 0, 0), thickness=self.region_thickness)
+        self.annotator.draw_region(reg_pts=self.reg_pts, color=region_color, thickness=self.region_thickness)
 
         for box, trk_id, cls in zip(self.boxes, self.trk_ids, self.clss):
             track = self.store_track_info(trk_id, box)