ultralytics 8.3.131__tar.gz → 8.3.132__tar.gz

{ultralytics-8.3.131/ultralytics.egg-info → ultralytics-8.3.132}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: ultralytics
-Version: 8.3.131
+Version: 8.3.132
 Summary: Ultralytics YOLO 🚀 for SOTA object detection, multi-object tracking, instance segmentation, pose estimation and image classification.
 Author-email: Glenn Jocher <glenn.jocher@ultralytics.com>, Jing Qiu <jing.qiu@ultralytics.com>
 Maintainer-email: Ultralytics <hello@ultralytics.com>

{ultralytics-8.3.131 → ultralytics-8.3.132}/ultralytics/__init__.py

@@ -1,6 +1,6 @@
 # Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license
 
-__version__ = "8.3.131"
+__version__ = "8.3.132"
 
 import os
 

ultralytics-8.3.132/ultralytics/cfg/datasets/HomeObjects-3K.yaml

@@ -0,0 +1,33 @@
+# Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license
+
+# HomeObjects-3K dataset by Ultralytics
+# Documentation: https://docs.ultralytics.com/datasets/detect/homeobjects-3k/
+# Example usage: yolo train data=HomeObjects-3K.yaml
+# parent
+# ├── ultralytics
+# └── datasets
+#     └── homeobjects-3K  ← downloads here (390 MB)
+
+# Train/val/test sets as 1) dir: path/to/imgs, 2) file: path/to/imgs.txt, or 3) list: [path/to/imgs1, path/to/imgs2, ..]
+path: ../datasets/homeobjects-3K # dataset root dir
+train: train/images # train images (relative to 'path') 2285 images
+val: valid/images # val images (relative to 'path') 404 images
+test: # test images (relative to 'path')
+
+# Classes
+names:
+  0: bed
+  1: sofa
+  2: chair
+  3: table
+  4: lamp
+  5: tv
+  6: laptop
+  7: wardrobe
+  8: window
+  9: door
+  10: potted plant
+  11: photo frame
+
+# Download script/URL (optional)
+download: https://github.com/ultralytics/assets/releases/download/v0.0.0/homeobjects-3K.zip

{ultralytics-8.3.131 → ultralytics-8.3.132}/ultralytics/data/dataset.py

@@ -39,7 +39,7 @@ from .utils import (
     verify_image_label,
 )
 
-# Ultralytics dataset *.cache version, >= 1.0.0 for YOLOv8
+# Ultralytics dataset *.cache version, >= 1.0.0 for Ultralytics YOLO models
 DATASET_CACHE_VERSION = "1.0.3"
 
 

{ultralytics-8.3.131 → ultralytics-8.3.132}/ultralytics/engine/trainer.py

@@ -77,8 +77,6 @@ class BaseTrainer:
         amp (bool): Flag to enable AMP (Automatic Mixed Precision).
         scaler (amp.GradScaler): Gradient scaler for AMP.
         data (str): Path to data.
-        trainset (torch.utils.data.Dataset): Training dataset.
-        testset (torch.utils.data.Dataset): Testing dataset.
         ema (nn.Module): EMA (Exponential Moving Average) of the model.
         resume (bool): Resume training from a checkpoint.
         lf (nn.Module): Loss function.
@@ -136,7 +134,8 @@ class BaseTrainer:
         # Model and Dataset
         self.model = check_model_file_from_stem(self.args.model)  # add suffix, i.e. yolo11n -> yolo11n.pt
         with torch_distributed_zero_first(LOCAL_RANK):  # avoid auto-downloading dataset multiple times
-            self.trainset, self.testset = self.get_dataset()
+            self.data = self.get_dataset()
+
         self.ema = None
 
         # Optimization utils init
@@ -289,11 +288,16 @@ class BaseTrainer:
 
         # Dataloaders
         batch_size = self.batch_size // max(world_size, 1)
-        self.train_loader = self.get_dataloader(self.trainset, batch_size=batch_size, rank=LOCAL_RANK, mode="train")
+        self.train_loader = self.get_dataloader(
+            self.data["train"], batch_size=batch_size, rank=LOCAL_RANK, mode="train"
+        )
         if RANK in {-1, 0}:
             # Note: When training DOTA dataset, double batch size could get OOM on images with >2000 objects.
             self.test_loader = self.get_dataloader(
-                self.testset, batch_size=batch_size if self.args.task == "obb" else batch_size * 2, rank=-1, mode="val"
+                self.data.get("val") or self.data.get("test"),
+                batch_size=batch_size if self.args.task == "obb" else batch_size * 2,
+                rank=-1,
+                mode="val",
             )
             self.validator = self.get_validator()
             metric_keys = self.validator.metrics.keys + self.label_loss_items(prefix="val")
@@ -569,7 +573,7 @@ class BaseTrainer:
         Get train and validation datasets from data dictionary.
 
         Returns:
-            (tuple): A tuple containing the training and validation/test datasets.
+            (dict): A dictionary containing the training/validation/test dataset and category names.
         """
         try:
             if self.args.task == "classify":
@@ -585,12 +589,11 @@ class BaseTrainer:
                     self.args.data = data["yaml_file"]  # for validating 'yolo train data=url.zip' usage
         except Exception as e:
             raise RuntimeError(emojis(f"Dataset '{clean_url(self.args.data)}' error ❌ {e}")) from e
-        self.data = data
         if self.args.single_cls:
             LOGGER.info("Overriding class names with single class.")
-            self.data["names"] = {0: "item"}
-            self.data["nc"] = 1
-        return data["train"], data.get("val") or data.get("test")
+            data["names"] = {0: "item"}
+            data["nc"] = 1
+        return data
 
     def setup_model(self):
         """

{ultralytics-8.3.131 → ultralytics-8.3.132}/ultralytics/models/yolo/detect/train.py

@@ -212,6 +212,6 @@ class DetectionTrainer(BaseTrainer):
         Returns:
             (int): Optimal batch size.
         """
-        train_dataset = self.build_dataset(self.trainset, mode="train", batch=16)
+        train_dataset = self.build_dataset(self.data["train"], mode="train", batch=16)
         max_num_obj = max(len(label["cls"]) for label in train_dataset.labels) * 4  # 4 for mosaic augmentation
         return super().auto_batch(max_num_obj)

{ultralytics-8.3.131 → ultralytics-8.3.132}/ultralytics/models/yolo/pose/train.py

@@ -137,3 +137,18 @@ class PoseTrainer(yolo.detect.DetectionTrainer):
     def plot_metrics(self):
         """Plots training/val metrics."""
         plot_results(file=self.csv, pose=True, on_plot=self.on_plot)  # save results.png
+
+    def get_dataset(self):
+        """
+        Retrieves the dataset and ensures it contains the required `kpt_shape` key.
+
+        Returns:
+            (dict): A dictionary containing the training/validation/test dataset and category names.
+
+        Raises:
+            KeyError: If the `kpt_shape` key is not present in the dataset.
+        """
+        data = super().get_dataset()
+        if "kpt_shape" not in data:
+            raise KeyError(f"No `kpt_shape` in the {self.args.data}. See https://docs.ultralytics.com/datasets/pose/")
+        return data

{ultralytics-8.3.131 → ultralytics-8.3.132}/ultralytics/models/yolo/world/train_world.py

@@ -137,12 +137,13 @@ class WorldTrainerFromScratch(WorldTrainer):
             for g in grounding_data:
                 assert isinstance(g, dict), f"Grounding data should be provided in dict format, but got {type(g)}"
             final_data[s] += grounding_data
+        data["val"] = data["val"][0]  # assign the first val dataset as currently only one validation set is supported
         # NOTE: to make training work properly, set `nc` and `names`
-        final_data["nc"] = data["val"][0]["nc"]
-        final_data["names"] = data["val"][0]["names"]
+        final_data["nc"] = data["val"]["nc"]
+        final_data["names"] = data["val"]["names"]
         # NOTE: add path with lvis path
-        final_data["path"] = data["val"][0]["path"]
-        final_data["channels"] = data["val"][0]["channels"]
+        final_data["path"] = data["val"]["path"]
+        final_data["channels"] = data["val"]["channels"]
         self.data = final_data
         if self.args.single_cls:  # consistent with base trainer
             LOGGER.info("Overriding class names with single class.")
@@ -154,7 +155,7 @@ class WorldTrainerFromScratch(WorldTrainer):
                 d["names"] = {0: "object"}
                 d["nc"] = 1
             self.training_data[d["train"]] = d
-        return final_data["train"], final_data["val"][0]
+        return final_data
 
     def plot_training_labels(self):
         """Do not plot labels for YOLO-World training."""

{ultralytics-8.3.131 → ultralytics-8.3.132}/ultralytics/models/yolo/yoloe/predict.py

@@ -1,6 +1,5 @@
 # Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license
 
-
 import numpy as np
 import torch
 

{ultralytics-8.3.131 → ultralytics-8.3.132}/ultralytics/models/yolo/yoloe/train_seg.py

@@ -1,6 +1,5 @@
 # Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license
 
-
 from copy import copy, deepcopy
 
 from ultralytics.models.yolo.segment import SegmentationTrainer

{ultralytics-8.3.131 → ultralytics-8.3.132}/ultralytics/nn/modules/block.py

@@ -76,11 +76,11 @@ class DFL(nn.Module):
 
 
 class Proto(nn.Module):
-    """YOLOv8 mask Proto module for segmentation models."""
+    """Ultralytics YOLO models mask Proto module for segmentation models."""
 
     def __init__(self, c1, c_=256, c2=32):
         """
-        Initialize the YOLOv8 mask Proto module with specified number of protos and masks.
+        Initialize the Ultralytics YOLO models mask Proto module with specified number of protos and masks.
 
         Args:
             c1 (int): Input channels.

{ultralytics-8.3.131 → ultralytics-8.3.132}/ultralytics/nn/tasks.py

@@ -281,10 +281,19 @@ class BaseModel(torch.nn.Module):
         """
         model = weights["model"] if isinstance(weights, dict) else weights  # torchvision models are not dicts
         csd = model.float().state_dict()  # checkpoint state_dict as FP32
-        csd = intersect_dicts(csd, self.state_dict())  # intersect
-        self.load_state_dict(csd, strict=False)  # load
+        updated_csd = intersect_dicts(csd, self.state_dict())  # intersect
+        self.load_state_dict(updated_csd, strict=False)  # load
+        len_updated_csd = len(updated_csd)
+        first_conv = "model.0.conv.weight"
+        if first_conv not in updated_csd:  # mostly used to boost multi-channel training
+            c1, c2, h, w = self.state_dict()[first_conv].shape
+            cc1, cc2, ch, cw = csd[first_conv].shape
+            if ch == h and cw == w:
+                c1, c2 = min(c1, cc1), min(c2, cc2)
+                self.state_dict()[first_conv][:c1, :c2] = csd[first_conv][:c1, :c2]
+                len_updated_csd += 1
         if verbose:
-            LOGGER.info(f"Transferred {len(csd)}/{len(self.model.state_dict())} items from pretrained weights")
+            LOGGER.info(f"Transferred {len_updated_csd}/{len(self.model.state_dict())} items from pretrained weights")
 
     def loss(self, batch, preds=None):
         """
@@ -458,7 +467,7 @@ class SegmentationModel(DetectionModel):
 
     def __init__(self, cfg="yolo11n-seg.yaml", ch=3, nc=None, verbose=True):
         """
-        Initialize YOLOv8 segmentation model with given config and parameters.
+        Initialize Ultralytics YOLO segmentation model with given config and parameters.
 
         Args:
             cfg (str | dict): Model configuration file path or dictionary.
@@ -478,7 +487,7 @@ class PoseModel(DetectionModel):
 
     def __init__(self, cfg="yolo11n-pose.yaml", ch=3, nc=None, data_kpt_shape=(None, None), verbose=True):
         """
-        Initialize YOLOv8 Pose model.
+        Initialize Ultralytics YOLO Pose model.
 
         Args:
             cfg (str | dict): Model configuration file path or dictionary.
@@ -517,7 +526,7 @@ class ClassificationModel(BaseModel):
 
     def _from_yaml(self, cfg, ch, nc, verbose):
         """
-        Set YOLOv8 model configurations and define the model architecture.
+        Set Ultralytics YOLO model configurations and define the model architecture.
 
         Args:
             cfg (str | dict): Model configuration file path or dictionary.

{ultralytics-8.3.131 → ultralytics-8.3.132}/ultralytics/solutions/heatmap.py

@@ -100,12 +100,11 @@ class Heatmap(ObjectCounter):
                 self.annotator.draw_region(reg_pts=self.region, color=(104, 0, 123), thickness=self.line_width * 2)
                 self.store_tracking_history(track_id, box)  # Store track history
                 self.store_classwise_counts(cls)  # Store classwise counts in dict
-                current_centroid = ((box[0] + box[2]) / 2, (box[1] + box[3]) / 2)
                 # Get previous position if available
                 prev_position = None
                 if len(self.track_history[track_id]) > 1:
                     prev_position = self.track_history[track_id][-2]
-                self.count_objects(current_centroid, track_id, prev_position, cls)  # Perform object counting
+                self.count_objects(self.track_history[track_id][-1], track_id, prev_position, cls)  # object counting
 
         plot_im = self.annotator.result()
         if self.region is not None:

{ultralytics-8.3.131 → ultralytics-8.3.132}/ultralytics/solutions/instance_segmentation.py

@@ -62,6 +62,7 @@ class InstanceSegmentation(BaseSolution):
             >>> print(summary)
         """
         self.extract_tracks(im0)  # Extract tracks (bounding boxes, classes, and masks)
+        self.masks = getattr(self.tracks[0], "masks", None)
 
         # Iterate over detected classes, track IDs, and segmentation masks
         if self.masks is None:

{ultralytics-8.3.131 → ultralytics-8.3.132}/ultralytics/solutions/object_blurrer.py

@@ -1,6 +1,5 @@
 # Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license
 
-
 import cv2
 
 from ultralytics.solutions.solutions import BaseSolution, SolutionAnnotator, SolutionResults

{ultralytics-8.3.131 → ultralytics-8.3.132}/ultralytics/solutions/object_counter.py

@@ -174,6 +174,10 @@ class ObjectCounter(BaseSolution):
         self.extract_tracks(im0)  # Extract tracks
         self.annotator = SolutionAnnotator(im0, line_width=self.line_width)  # Initialize annotator
 
+        is_obb = getattr(self.tracks[0], "obb", None) is not None  # True if OBB results exist
+        if is_obb:
+            self.boxes = self.track_data.xyxyxyxy.reshape(-1, 4, 2).cpu()
+
         self.annotator.draw_region(
             reg_pts=self.region, color=(104, 0, 123), thickness=self.line_width * 2
         )  # Draw region
@@ -181,16 +185,17 @@ class ObjectCounter(BaseSolution):
         # Iterate over bounding boxes, track ids and classes index
         for box, track_id, cls, conf in zip(self.boxes, self.track_ids, self.clss, self.confs):
             # Draw bounding box and counting region
-            self.annotator.box_label(box, label=self.adjust_box_label(cls, conf, track_id), color=colors(cls, True))
-            self.store_tracking_history(track_id, box)  # Store track history
+            self.annotator.box_label(
+                box, label=self.adjust_box_label(cls, conf, track_id), color=colors(cls, True), rotated=is_obb
+            )
+            self.store_tracking_history(track_id, box, is_obb=is_obb)  # Store track history
             self.store_classwise_counts(cls)  # Store classwise counts in dict
 
-            current_centroid = ((box[0] + box[2]) / 2, (box[1] + box[3]) / 2)
             # Store previous position of track for object counting
             prev_position = None
             if len(self.track_history[track_id]) > 1:
                 prev_position = self.track_history[track_id][-2]
-            self.count_objects(current_centroid, track_id, prev_position, cls)  # Perform object counting
+            self.count_objects(self.track_history[track_id][-1], track_id, prev_position, cls)  # object counting
 
         plot_im = self.annotator.result()
         self.display_counts(plot_im)  # Display the counts on the frame

{ultralytics-8.3.131 → ultralytics-8.3.132}/ultralytics/solutions/solutions.py

@@ -139,10 +139,6 @@ class BaseSolution:
         self.tracks = self.model.track(source=im0, persist=True, classes=self.classes, **self.track_add_args)
         self.track_data = self.tracks[0].obb or self.tracks[0].boxes  # Extract tracks for OBB or object detection
 
-        self.masks = (
-            self.tracks[0].masks if hasattr(self.tracks[0], "masks") and self.tracks[0].masks is not None else None
-        )
-
         if self.track_data and self.track_data.id is not None:
             self.boxes = self.track_data.xyxy.cpu()
             self.clss = self.track_data.cls.cpu().tolist()
@@ -152,7 +148,7 @@ class BaseSolution:
             self.LOGGER.warning("no tracks found!")
             self.boxes, self.clss, self.track_ids, self.confs = [], [], [], []
 
-    def store_tracking_history(self, track_id, box):
+    def store_tracking_history(self, track_id, box, is_obb=False):
         """
         Stores the tracking history of an object.
 
@@ -162,6 +158,7 @@ class BaseSolution:
         Args:
             track_id (int): The unique identifier for the tracked object.
             box (List[float]): The bounding box coordinates of the object in the format [x1, y1, x2, y2].
+            is_obb (bool): True if OBB model is used (applies to object counting only).
 
         Examples:
             >>> solution = BaseSolution()
@@ -169,7 +166,7 @@ class BaseSolution:
         """
         # Store tracking history
         self.track_line = self.track_history[track_id]
-        self.track_line.append(((box[0] + box[2]) / 2, (box[1] + box[3]) / 2))
+        self.track_line.append(tuple(box.mean(dim=0)) if is_obb else (box[:4:2].mean(), box[1:4:2].mean()))
         if len(self.track_line) > 30:
             self.track_line.pop(0)
 

{ultralytics-8.3.131 → ultralytics-8.3.132}/ultralytics/solutions/vision_eye.py

@@ -1,6 +1,5 @@
 # Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license
 
-
 from ultralytics.solutions.solutions import BaseSolution, SolutionAnnotator, SolutionResults
 from ultralytics.utils.plotting import colors
 

{ultralytics-8.3.131 → ultralytics-8.3.132}/ultralytics/utils/autodevice.py

@@ -1,6 +1,5 @@
 # Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license
 
-
 from ultralytics.utils import LOGGER
 from ultralytics.utils.checks import check_requirements
 

{ultralytics-8.3.131 → ultralytics-8.3.132}/ultralytics/utils/callbacks/neptune.py

@@ -1,6 +1,5 @@
 # Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license
 
-
 from ultralytics.utils import LOGGER, SETTINGS, TESTS_RUNNING
 
 try:

{ultralytics-8.3.131 → ultralytics-8.3.132}/ultralytics/utils/checks.py

@@ -343,7 +343,7 @@ def check_python(minimum: str = "3.8.0", hard: bool = True, verbose: bool = Fals
 @TryExcept()
 def check_requirements(requirements=ROOT.parent / "requirements.txt", exclude=(), install=True, cmds=""):
     """
-    Check if installed dependencies meet YOLOv8 requirements and attempt to auto-update if needed.
+    Check if installed dependencies meet Ultralytics YOLO models requirements and attempt to auto-update if needed.
 
     Args:
         requirements (Union[Path, str, List[str]]): Path to a requirements.txt file, a single package requirement as a

{ultralytics-8.3.131 → ultralytics-8.3.132}/ultralytics/utils/metrics.py

@@ -665,7 +665,7 @@ def ap_per_class(
 
 class Metric(SimpleClass):
     """
-    Class for computing evaluation metrics for YOLOv8 model.
+    Class for computing evaluation metrics for Ultralytics YOLO models.
 
     Attributes:
         p (list): Precision for each class. Shape: (nc,).

{ultralytics-8.3.131 → ultralytics-8.3.132/ultralytics.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: ultralytics
-Version: 8.3.131
+Version: 8.3.132
 Summary: Ultralytics YOLO 🚀 for SOTA object detection, multi-object tracking, instance segmentation, pose estimation and image classification.
 Author-email: Glenn Jocher <glenn.jocher@ultralytics.com>, Jing Qiu <jing.qiu@ultralytics.com>
 Maintainer-email: Ultralytics <hello@ultralytics.com>

{ultralytics-8.3.131 → ultralytics-8.3.132}/ultralytics.egg-info/SOURCES.txt

@@ -32,6 +32,7 @@ ultralytics/cfg/datasets/Argoverse.yaml
 ultralytics/cfg/datasets/DOTAv1.5.yaml
 ultralytics/cfg/datasets/DOTAv1.yaml
 ultralytics/cfg/datasets/GlobalWheat2020.yaml
+ultralytics/cfg/datasets/HomeObjects-3K.yaml
 ultralytics/cfg/datasets/ImageNet.yaml
 ultralytics/cfg/datasets/Objects365.yaml
 ultralytics/cfg/datasets/SKU-110K.yaml