ultralytics 8.3.69__tar.gz → 8.3.71__tar.gz

{ultralytics-8.3.69/ultralytics.egg-info → ultralytics-8.3.71}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.2
 Name: ultralytics
-Version: 8.3.69
+Version: 8.3.71
 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>
@@ -32,8 +32,7 @@ Classifier: Operating System :: Microsoft :: Windows
 Requires-Python: >=3.8
 Description-Content-Type: text/markdown
 License-File: LICENSE
-Requires-Dist: numpy>=1.23.0
-Requires-Dist: numpy<2.0.0; sys_platform == "darwin"
+Requires-Dist: numpy<=2.1.1,>=1.23.0
 Requires-Dist: matplotlib>=3.3.0
 Requires-Dist: opencv-python>=4.6.0
 Requires-Dist: pillow>=7.1.2
@@ -55,10 +54,11 @@ Requires-Dist: pytest; extra == "dev"
 Requires-Dist: pytest-cov; extra == "dev"
 Requires-Dist: coverage[toml]; extra == "dev"
 Requires-Dist: mkdocs>=1.6.0; extra == "dev"
+Requires-Dist: beautifulsoup4<=4.12.3; extra == "dev"
 Requires-Dist: mkdocs-material>=9.5.9; extra == "dev"
 Requires-Dist: mkdocstrings[python]; extra == "dev"
 Requires-Dist: mkdocs-redirects; extra == "dev"
-Requires-Dist: mkdocs-ultralytics-plugin>=0.1.8; extra == "dev"
+Requires-Dist: mkdocs-ultralytics-plugin>=0.1.16; extra == "dev"
 Requires-Dist: mkdocs-macros-plugin>=1.0.5; extra == "dev"
 Provides-Extra: export
 Requires-Dist: onnx>=1.12.0; extra == "export"

{ultralytics-8.3.69 → ultralytics-8.3.71}/pyproject.toml

@@ -61,8 +61,7 @@ classifiers = [
 
 # Required dependencies ------------------------------------------------------------------------------------------------
 dependencies = [
-    "numpy>=1.23.0",
-    "numpy<2.0.0; sys_platform == 'darwin'", # macOS OpenVINO errors https://github.com/ultralytics/ultralytics/pull/17221
+    "numpy>=1.23.0,<=2.1.1", # OpenVINO and TFLite errors on '--slow' CI Tests https://github.com/ultralytics/ultralytics/pull/18943
     "matplotlib>=3.3.0",
     "opencv-python>=4.6.0",
     "pillow>=7.1.2",
@@ -88,10 +87,11 @@ dev = [
     "pytest-cov",
     "coverage[toml]",
     "mkdocs>=1.6.0",
+    "beautifulsoup4<=4.12.3",  # For docs https://github.com/ultralytics/ultralytics/pull/19067
     "mkdocs-material>=9.5.9",
     "mkdocstrings[python]",
     "mkdocs-redirects", # 301 redirects
-    "mkdocs-ultralytics-plugin>=0.1.8", # for meta descriptions and images, dates and authors
+    "mkdocs-ultralytics-plugin>=0.1.16", # for meta descriptions and images, dates and authors
     "mkdocs-macros-plugin>=1.0.5"  # duplicating content (i.e. export tables) in multiple places
 ]
 export = [

{ultralytics-8.3.69 → ultralytics-8.3.71}/ultralytics/__init__.py

@@ -1,6 +1,6 @@
 # Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license
 
-__version__ = "8.3.69"
+__version__ = "8.3.71"
 
 import os
 

{ultralytics-8.3.69 → ultralytics-8.3.71}/ultralytics/cfg/models/v6/yolov6.yaml

@@ -6,7 +6,7 @@
 
 # Parameters
 nc: 80 # number of classes
-activation: nn.ReLU() # (optional) model default activation function
+activation: torch.nn.ReLU() # (optional) model default activation function
 scales: # model compound scaling constants, i.e. 'model=yolov6n.yaml' will call yolov8.yaml with scale 'n'
   # [depth, width, max_channels]
   n: [0.33, 0.25, 1024]

{ultralytics-8.3.69 → ultralytics-8.3.71}/ultralytics/engine/model.py

@@ -11,7 +11,7 @@ from PIL import Image
 from ultralytics.cfg import TASK2DATA, get_cfg, get_save_dir
 from ultralytics.engine.results import Results
 from ultralytics.hub import HUB_WEB_ROOT, HUBTrainingSession
-from ultralytics.nn.tasks import attempt_load_one_weight, guess_model_task, nn, yaml_model_load
+from ultralytics.nn.tasks import attempt_load_one_weight, guess_model_task, yaml_model_load
 from ultralytics.utils import (
     ARGV,
     ASSETS,
@@ -26,7 +26,7 @@ from ultralytics.utils import (
 )
 
 
-class Model(nn.Module):
+class Model(torch.nn.Module):
     """
     A base class for implementing YOLO models, unifying APIs across different model types.
 
@@ -37,7 +37,7 @@ class Model(nn.Module):
     Attributes:
         callbacks (Dict): A dictionary of callback functions for various events during model operations.
         predictor (BasePredictor): The predictor object used for making predictions.
-        model (nn.Module): The underlying PyTorch model.
+        model (torch.nn.Module): The underlying PyTorch model.
         trainer (BaseTrainer): The trainer object used for training the model.
         ckpt (Dict): The checkpoint data if the model is loaded from a *.pt file.
         cfg (str): The configuration of the model if loaded from a *.yaml file.
@@ -317,7 +317,7 @@ class Model(nn.Module):
             >>> model._check_is_pytorch_model()  # Raises TypeError
         """
         pt_str = isinstance(self.model, (str, Path)) and Path(self.model).suffix == ".pt"
-        pt_module = isinstance(self.model, nn.Module)
+        pt_module = isinstance(self.model, torch.nn.Module)
         if not (pt_module or pt_str):
             raise TypeError(
                 f"model='{self.model}' should be a *.pt PyTorch model to run this method, but is a different format. "
@@ -405,7 +405,7 @@ class Model(nn.Module):
         from ultralytics import __version__
 
         updates = {
-            "model": deepcopy(self.model).half() if isinstance(self.model, nn.Module) else self.model,
+            "model": deepcopy(self.model).half() if isinstance(self.model, torch.nn.Module) else self.model,
             "date": datetime.now().isoformat(),
             "version": __version__,
             "license": "AGPL-3.0 License (https://ultralytics.com/license)",
@@ -452,7 +452,7 @@ class Model(nn.Module):
         performs both convolution and normalization in one step.
 
         Raises:
-            TypeError: If the model is not a PyTorch nn.Module.
+            TypeError: If the model is not a PyTorch torch.nn.Module.
 
         Examples:
             >>> model = Model("yolo11n.pt")
@@ -661,6 +661,7 @@ class Model(nn.Module):
                 - int8 (bool): Whether to use int8 precision mode.
                 - device (str): Device to run the benchmark on (e.g., 'cpu', 'cuda').
                 - verbose (bool): Whether to print detailed benchmark information.
+                - format (str): Export format name for specific benchmarking
 
         Returns:
             (Dict): A dictionary containing the results of the benchmarking process, including metrics for
@@ -686,7 +687,8 @@ class Model(nn.Module):
             half=args["half"],
             int8=args["int8"],
             device=args["device"],
-            verbose=kwargs.get("verbose"),
+            verbose=kwargs.get("verbose", False),
+            format=kwargs.get("format", ""),
         )
 
     def export(
@@ -919,13 +921,13 @@ class Model(nn.Module):
         Retrieves the device on which the model's parameters are allocated.
 
         This property determines the device (CPU or GPU) where the model's parameters are currently stored. It is
-        applicable only to models that are instances of nn.Module.
+        applicable only to models that are instances of torch.nn.Module.
 
         Returns:
             (torch.device): The device (CPU/GPU) of the model.
 
         Raises:
-            AttributeError: If the model is not a PyTorch nn.Module instance.
+            AttributeError: If the model is not a torch.nn.Module instance.
 
         Examples:
             >>> model = YOLO("yolo11n.pt")
@@ -935,7 +937,7 @@ class Model(nn.Module):
             >>> print(model.device)
             device(type='cpu')
         """
-        return next(self.model.parameters()).device if isinstance(self.model, nn.Module) else None
+        return next(self.model.parameters()).device if isinstance(self.model, torch.nn.Module) else None
 
     @property
     def transforms(self):

{ultralytics-8.3.69 → ultralytics-8.3.71}/ultralytics/engine/results.py

@@ -494,8 +494,8 @@ class Results(SimpleClass):
         Examples:
             >>> results = model("image.jpg")
             >>> for result in results:
-            ...     im = result.plot()
-            ...     im.show()
+            >>>     im = result.plot()
+            >>>     im.show()
         """
         assert color_mode in {"instance", "class"}, f"Expected color_mode='instance' or 'class', not {color_mode}."
         if img is None and isinstance(self.orig_img, torch.Tensor):
@@ -600,7 +600,7 @@ class Results(SimpleClass):
             >>> results = model("path/to/image.jpg")
             >>> results[0].show()  # Display the first result
             >>> for result in results:
-            ...     result.show()  # Display all results
+            >>>     result.show()  # Display all results
         """
         self.plot(show=True, *args, **kwargs)
 
@@ -620,10 +620,10 @@ class Results(SimpleClass):
         Examples:
             >>> results = model("path/to/image.jpg")
             >>> for result in results:
-            ...     result.save("annotated_image.jpg")
+            >>>     result.save("annotated_image.jpg")
             >>> # Or with custom plot arguments
             >>> for result in results:
-            ...     result.save("annotated_image.jpg", conf=False, line_width=2)
+            >>>     result.save("annotated_image.jpg", conf=False, line_width=2)
         """
         if not filename:
             filename = f"results_{Path(self.path).name}"
@@ -644,7 +644,7 @@ class Results(SimpleClass):
         Examples:
             >>> results = model("path/to/image.jpg")
             >>> for result in results:
-            ...     print(result.verbose())
+            >>>     print(result.verbose())
             2 persons, 1 car, 3 traffic lights,
             dog 0.92, cat 0.78, horse 0.64,
 
@@ -681,7 +681,7 @@ class Results(SimpleClass):
             >>> model = YOLO("yolo11n.pt")
             >>> results = model("path/to/image.jpg")
             >>> for result in results:
-            ...     result.save_txt("output.txt")
+            >>>     result.save_txt("output.txt")
 
         Notes:
             - The file will contain one line per detection or classification with the following structure:
@@ -740,7 +740,7 @@ class Results(SimpleClass):
         Examples:
             >>> results = model("path/to/image.jpg")
             >>> for result in results:
-            ...     result.save_crop(save_dir="path/to/crops", file_name="detection")
+            >>>     result.save_crop(save_dir="path/to/crops", file_name="detection")
         """
         if self.probs is not None:
             LOGGER.warning("WARNING ⚠️ Classify task do not support `save_crop`.")
@@ -776,8 +776,9 @@ class Results(SimpleClass):
 
         Examples:
             >>> results = model("image.jpg")
-            >>> summary = results[0].summary()
-            >>> print(summary)
+            >>> for result in results:
+            >>>     summary = result.summary()
+            >>>     print(summary)
         """
         # Create list of detection dictionaries
         results = []
@@ -839,8 +840,9 @@ class Results(SimpleClass):
 
         Examples:
             >>> results = model("path/to/image.jpg")
-            >>> df_result = results[0].to_df()
-            >>> print(df_result)
+            >>> for result in results:
+            >>>     df_result = result.to_df()
+            >>>     print(df_result)
         """
         import pandas as pd  # scope for faster 'import ultralytics'
 
@@ -867,8 +869,9 @@ class Results(SimpleClass):
 
         Examples:
             >>> results = model("path/to/image.jpg")
-            >>> csv_result = results[0].to_csv()
-            >>> print(csv_result)
+            >>> for result in results:
+            >>>     csv_result = result.to_csv()
+            >>>     print(csv_result)
         """
         return self.to_df(normalize=normalize, decimals=decimals).to_csv(*args, **kwargs)
 
@@ -892,8 +895,9 @@ class Results(SimpleClass):
 
         Examples:
             >>> results = model("path/to/image.jpg")
-            >>> xml_result = results[0].to_xml()
-            >>> print(xml_result)
+            >>> for result in results:
+            >>>     xml_result = result.to_xml()
+            >>>     print(xml_result)
         """
         check_requirements("lxml")
         df = self.to_df(normalize=normalize, decimals=decimals)
@@ -922,8 +926,9 @@ class Results(SimpleClass):
 
         Examples:
             >>> results = model("path/to/image.jpg")
-            >>> json_result = results[0].to_json()
-            >>> print(json_result)
+            >>> for result in results:
+            >>>     json_result = result.to_json()
+            >>>     print(json_result)
 
         Notes:
             - For classification tasks, the JSON will contain class probabilities instead of bounding boxes.
@@ -954,8 +959,8 @@ class Results(SimpleClass):
 
         Examples:
             >>> results = model("path/to/image.jpg")
-            >>> results[0].to_sql()
-            >>> print("SQL data written successfully.")
+            >>> for result in results:
+            >>>     result.to_sql()
         """
         import json
         import sqlite3

{ultralytics-8.3.69 → ultralytics-8.3.71}/ultralytics/models/sam/modules/sam.py

@@ -426,8 +426,7 @@ class SAM2Model(torch.nn.Module):
                 high_res_masks: Tensor of shape (B, 1, H*16, W*16) with the best high-resolution mask.
                 obj_ptr: Tensor of shape (B, C) with object pointer vector for the output mask.
                 object_score_logits: Tensor of shape (B) with object score logits.
-
-            Where M is 3 if multimask_output=True, and 1 if multimask_output=False.
+                Where M is 3 if multimask_output=True, and 1 if multimask_output=False.
 
         Examples:
             >>> backbone_features = torch.rand(1, 256, 32, 32)

{ultralytics-8.3.69 → ultralytics-8.3.71}/ultralytics/models/yolo/pose/val.py

@@ -158,7 +158,7 @@ class PoseValidator(DetectionValidator):
             gt_kpts (torch.Tensor | None): Optional tensor with shape (N, 51) representing ground truth keypoints.
 
         Returns:
-            torch.Tensor: A tensor with shape (N, 10) representing the correct prediction matrix for 10 IoU levels,
+            (torch.Tensor): A tensor with shape (N, 10) representing the correct prediction matrix for 10 IoU levels,
                 where N is the number of detections.
 
         Example:

{ultralytics-8.3.69 → ultralytics-8.3.71}/ultralytics/nn/autobackend.py

@@ -293,6 +293,12 @@ class AutoBackend(nn.Module):
                 except UnicodeDecodeError:
                     f.seek(0)  # engine file may lack embedded Ultralytics metadata
                 model = runtime.deserialize_cuda_engine(f.read())  # read engine
+                if "dla" in str(device.type):
+                    dla_core = int(device.type.split(":")[1])
+                    assert dla_core in {0, 1}, (
+                        "Expected device type for inference in DLA is 'dla:0' or 'dla:1', but received '{device.type}'"
+                    )
+                    runtime.DLA_core = dla_core
 
             # Model context
             try:
@@ -774,7 +780,7 @@ class AutoBackend(nn.Module):
         saved_model, pb, tflite, edgetpu, tfjs, ncnn or paddle.
 
         Args:
-            p: path to the model file. Defaults to path/to/model.pt
+            p (str): path to the model file. Defaults to path/to/model.pt
 
         Examples:
             >>> model = AutoBackend(weights="path/to/model.onnx")

{ultralytics-8.3.69 → ultralytics-8.3.71}/ultralytics/nn/tasks.py

@@ -9,7 +9,6 @@ from pathlib import Path
 
 import thop
 import torch
-import torch.nn as nn
 
 from ultralytics.nn.modules import (
     AIFI,
@@ -88,7 +87,7 @@ from ultralytics.utils.torch_utils import (
 )
 
 
-class BaseModel(nn.Module):
+class BaseModel(torch.nn.Module):
     """The BaseModel class serves as a base class for all the models in the Ultralytics YOLO family."""
 
     def forward(self, x, *args, **kwargs):
@@ -151,7 +150,7 @@ class BaseModel(nn.Module):
             if visualize:
                 feature_visualization(x, m.type, m.i, save_dir=visualize)
             if embed and m.i in embed:
-                embeddings.append(nn.functional.adaptive_avg_pool2d(x, (1, 1)).squeeze(-1).squeeze(-1))  # flatten
+                embeddings.append(torch.nn.functional.adaptive_avg_pool2d(x, (1, 1)).squeeze(-1).squeeze(-1))  # flatten
                 if m.i == max(embed):
                     return torch.unbind(torch.cat(embeddings, 1), dim=0)
         return x
@@ -170,12 +169,9 @@ class BaseModel(nn.Module):
         the provided list.
 
         Args:
-            m (nn.Module): The layer to be profiled.
+            m (torch.nn.Module): The layer to be profiled.
             x (torch.Tensor): The input data to the layer.
             dt (list): A list to store the computation time of the layer.
-
-        Returns:
-            None
         """
         c = m == self.model[-1] and isinstance(x, list)  # is final layer list, copy input as inplace fix
         flops = thop.profile(m, inputs=[x.copy() if c else x], verbose=False)[0] / 1e9 * 2 if thop else 0  # GFLOPs
@@ -195,7 +191,7 @@ class BaseModel(nn.Module):
         computation efficiency.
 
         Returns:
-            (nn.Module): The fused model is returned.
+            (torch.nn.Module): The fused model is returned.
         """
         if not self.is_fused():
             for m in self.model.modules():
@@ -229,7 +225,7 @@ class BaseModel(nn.Module):
         Returns:
             (bool): True if the number of BatchNorm layers in the model is less than the threshold, False otherwise.
         """
-        bn = tuple(v for k, v in nn.__dict__.items() if "Norm" in k)  # normalization layers, i.e. BatchNorm2d()
+        bn = tuple(v for k, v in torch.nn.__dict__.items() if "Norm" in k)  # normalization layers, i.e. BatchNorm2d()
         return sum(isinstance(v, bn) for v in self.modules()) < thresh  # True if < 'thresh' BatchNorm layers in model
 
     def info(self, detailed=False, verbose=True, imgsz=640):
@@ -304,7 +300,7 @@ class DetectionModel(BaseModel):
         self.yaml = cfg if isinstance(cfg, dict) else yaml_model_load(cfg)  # cfg dict
         if self.yaml["backbone"][0][2] == "Silence":
             LOGGER.warning(
-                "WARNING ⚠️ YOLOv9 `Silence` module is deprecated in favor of nn.Identity. "
+                "WARNING ⚠️ YOLOv9 `Silence` module is deprecated in favor of torch.nn.Identity. "
                 "Please delete local *.pt file and re-download the latest model checkpoint."
             )
             self.yaml["backbone"][0][2] = "nn.Identity"
@@ -458,20 +454,22 @@ class ClassificationModel(BaseModel):
         name, m = list((model.model if hasattr(model, "model") else model).named_children())[-1]  # last module
         if isinstance(m, Classify):  # YOLO Classify() head
             if m.linear.out_features != nc:
-                m.linear = nn.Linear(m.linear.in_features, nc)
-        elif isinstance(m, nn.Linear):  # ResNet, EfficientNet
+                m.linear = torch.nn.Linear(m.linear.in_features, nc)
+        elif isinstance(m, torch.nn.Linear):  # ResNet, EfficientNet
             if m.out_features != nc:
-                setattr(model, name, nn.Linear(m.in_features, nc))
-        elif isinstance(m, nn.Sequential):
+                setattr(model, name, torch.nn.Linear(m.in_features, nc))
+        elif isinstance(m, torch.nn.Sequential):
             types = [type(x) for x in m]
-            if nn.Linear in types:
-                i = len(types) - 1 - types[::-1].index(nn.Linear)  # last nn.Linear index
+            if torch.nn.Linear in types:
+                i = len(types) - 1 - types[::-1].index(torch.nn.Linear)  # last torch.nn.Linear index
                 if m[i].out_features != nc:
-                    m[i] = nn.Linear(m[i].in_features, nc)
-            elif nn.Conv2d in types:
-                i = len(types) - 1 - types[::-1].index(nn.Conv2d)  # last nn.Conv2d index
+                    m[i] = torch.nn.Linear(m[i].in_features, nc)
+            elif torch.nn.Conv2d in types:
+                i = len(types) - 1 - types[::-1].index(torch.nn.Conv2d)  # last torch.nn.Conv2d index
                 if m[i].out_channels != nc:
-                    m[i] = nn.Conv2d(m[i].in_channels, nc, m[i].kernel_size, m[i].stride, bias=m[i].bias is not None)
+                    m[i] = torch.nn.Conv2d(
+                        m[i].in_channels, nc, m[i].kernel_size, m[i].stride, bias=m[i].bias is not None
+                    )
 
     def init_criterion(self):
         """Initialize the loss criterion for the ClassificationModel."""
@@ -587,7 +585,7 @@ class RTDETRDetectionModel(DetectionModel):
             if visualize:
                 feature_visualization(x, m.type, m.i, save_dir=visualize)
             if embed and m.i in embed:
-                embeddings.append(nn.functional.adaptive_avg_pool2d(x, (1, 1)).squeeze(-1).squeeze(-1))  # flatten
+                embeddings.append(torch.nn.functional.adaptive_avg_pool2d(x, (1, 1)).squeeze(-1).squeeze(-1))  # flatten
                 if m.i == max(embed):
                     return torch.unbind(torch.cat(embeddings, 1), dim=0)
         head = self.model[-1]
@@ -663,7 +661,7 @@ class WorldModel(DetectionModel):
             if visualize:
                 feature_visualization(x, m.type, m.i, save_dir=visualize)
             if embed and m.i in embed:
-                embeddings.append(nn.functional.adaptive_avg_pool2d(x, (1, 1)).squeeze(-1).squeeze(-1))  # flatten
+                embeddings.append(torch.nn.functional.adaptive_avg_pool2d(x, (1, 1)).squeeze(-1).squeeze(-1))  # flatten
                 if m.i == max(embed):
                     return torch.unbind(torch.cat(embeddings, 1), dim=0)
         return x
@@ -684,7 +682,7 @@ class WorldModel(DetectionModel):
         return self.criterion(preds, batch)
 
 
-class Ensemble(nn.ModuleList):
+class Ensemble(torch.nn.ModuleList):
     """Ensemble of models."""
 
     def __init__(self):
@@ -887,7 +885,7 @@ 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"):
+        elif isinstance(m, torch.nn.Upsample) and not hasattr(m, "recompute_scale_factor"):
             m.recompute_scale_factor = None  # torch 1.11.0 compatibility
 
     # Return model
@@ -922,7 +920,7 @@ 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"):
+        elif isinstance(m, torch.nn.Upsample) and not hasattr(m, "recompute_scale_factor"):
             m.recompute_scale_factor = None  # torch 1.11.0 compatibility
 
     # Return model and ckpt
@@ -946,7 +944,7 @@ def parse_model(d, ch, verbose=True):  # model_dict, input_channels(3)
         depth, width, max_channels = scales[scale]
 
     if act:
-        Conv.default_act = eval(act)  # redefine default activation, i.e. Conv.default_act = nn.SiLU()
+        Conv.default_act = eval(act)  # redefine default activation, i.e. Conv.default_act = torch.nn.SiLU()
         if verbose:
             LOGGER.info(f"{colorstr('activation:')} {act}")  # print
 
@@ -982,7 +980,7 @@ def parse_model(d, ch, verbose=True):  # model_dict, input_channels(3)
             C3,
             C3TR,
             C3Ghost,
-            nn.ConvTranspose2d,
+            torch.nn.ConvTranspose2d,
             DWConvTranspose2d,
             C3x,
             RepC3,
@@ -1048,7 +1046,7 @@ def parse_model(d, ch, verbose=True):  # model_dict, input_channels(3)
                 n = 1
         elif m is ResNetLayer:
             c2 = args[1] if args[3] else args[1] * 4
-        elif m is nn.BatchNorm2d:
+        elif m is torch.nn.BatchNorm2d:
             args = [ch[f]]
         elif m is Concat:
             c2 = sum(ch[x] for x in f)
@@ -1073,7 +1071,7 @@ def parse_model(d, ch, verbose=True):  # model_dict, input_channels(3)
         else:
             c2 = ch[f]
 
-        m_ = nn.Sequential(*(m(*args) for _ in range(n))) if n > 1 else m(*args)  # module
+        m_ = torch.nn.Sequential(*(m(*args) for _ in range(n))) if n > 1 else m(*args)  # module
         t = str(m)[8:-2].replace("__main__.", "")  # module type
         m_.np = sum(x.numel() for x in m_.parameters())  # number params
         m_.i, m_.f, m_.type = i, f, t  # attach index, 'from' index, type
@@ -1084,7 +1082,7 @@ def parse_model(d, ch, verbose=True):  # model_dict, input_channels(3)
         if i == 0:
             ch = []
         ch.append(c2)
-    return nn.Sequential(*layers), sorted(save)
+    return torch.nn.Sequential(*layers), sorted(save)
 
 
 def yaml_model_load(path):
@@ -1126,7 +1124,7 @@ def guess_model_task(model):
     Guess the task of a PyTorch model from its architecture or configuration.
 
     Args:
-        model (nn.Module | dict): PyTorch model or model configuration in YAML format.
+        model (torch.nn.Module | dict): PyTorch model or model configuration in YAML format.
 
     Returns:
         (str): Task of the model ('detect', 'segment', 'classify', 'pose').
@@ -1154,7 +1152,7 @@ def guess_model_task(model):
         with contextlib.suppress(Exception):
             return cfg2task(model)
     # Guess from PyTorch model
-    if isinstance(model, nn.Module):  # PyTorch model
+    if isinstance(model, torch.nn.Module):  # PyTorch model
         for x in "model.args", "model.model.args", "model.model.model.args":
             with contextlib.suppress(Exception):
                 return eval(x)["task"]

{ultralytics-8.3.69 → ultralytics-8.3.71}/ultralytics/utils/benchmarks.py

@@ -57,6 +57,7 @@ def benchmark(
     device="cpu",
     verbose=False,
     eps=1e-3,
+    format="",
 ):
     """
     Benchmark a YOLO model across different formats for speed and accuracy.
@@ -70,6 +71,7 @@ def benchmark(
         device (str): Device to run the benchmark on, either 'cpu' or 'cuda'.
         verbose (bool | float): If True or a float, assert benchmarks pass with given metric.
         eps (float): Epsilon value for divide by zero prevention.
+        format (str): Export format for benchmarking. If not supplied all formats are benchmarked.
 
     Returns:
         (pandas.DataFrame): A pandas DataFrame with benchmark results for each format, including file size, metric,
@@ -94,9 +96,17 @@ def benchmark(
 
     y = []
     t0 = time.time()
+
+    format_arg = format.lower()
+    if format_arg:
+        formats = frozenset(export_formats()["Argument"])
+        assert format in formats, f"Expected format to be one of {formats}, but got '{format_arg}'."
     for i, (name, format, suffix, cpu, gpu, _) in enumerate(zip(*export_formats().values())):
         emoji, filename = "❌", None  # export defaults
         try:
+            if format_arg and format_arg != format:
+                continue
+
             # Checks
             if i == 7:  # TF GraphDef
                 assert model.task != "obb", "TensorFlow GraphDef not supported for OBB task"
@@ -155,10 +165,10 @@ def benchmark(
 
             # Validate
             data = data or TASK2DATA[model.task]  # task to dataset, i.e. coco8.yaml for task=detect
-            key = TASK2METRIC[model.task]  # task to metric, i.e. metrics/mAP50-95(B) for task=detect
             results = exported_model.val(
                 data=data, batch=1, imgsz=imgsz, plots=False, device=device, half=half, int8=int8, verbose=False
             )
+            key = TASK2METRIC[model.task]  # task to metric, i.e. metrics/mAP50-95(B) for task=detect
             metric, speed = results.results_dict[key], results.speed["inference"]
             fps = round(1000 / (speed + eps), 2)  # frames per second
             y.append([name, "✅", round(file_size(filename), 1), round(metric, 4), round(speed, 2), fps])

{ultralytics-8.3.69 → ultralytics-8.3.71}/ultralytics/utils/checks.py

@@ -433,8 +433,8 @@ def check_torchvision():
     The compatibility table is a dictionary where the keys are PyTorch versions and the values are lists of compatible
     Torchvision versions.
     """
-    # Compatibility table
     compatibility_table = {
+        "2.6": ["0.21"],
         "2.5": ["0.20"],
         "2.4": ["0.19"],
         "2.3": ["0.18"],
@@ -445,7 +445,7 @@ def check_torchvision():
         "1.12": ["0.13"],
     }
 
-    # Extract only the major and minor versions
+    # Check major and minor versions
     v_torch = ".".join(torch.__version__.split("+")[0].split(".")[:2])
     if v_torch in compatibility_table:
         compatible_versions = compatibility_table[v_torch]

{ultralytics-8.3.69 → ultralytics-8.3.71/ultralytics.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.2
 Name: ultralytics
-Version: 8.3.69
+Version: 8.3.71
 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>
@@ -32,8 +32,7 @@ Classifier: Operating System :: Microsoft :: Windows
 Requires-Python: >=3.8
 Description-Content-Type: text/markdown
 License-File: LICENSE
-Requires-Dist: numpy>=1.23.0
-Requires-Dist: numpy<2.0.0; sys_platform == "darwin"
+Requires-Dist: numpy<=2.1.1,>=1.23.0
 Requires-Dist: matplotlib>=3.3.0
 Requires-Dist: opencv-python>=4.6.0
 Requires-Dist: pillow>=7.1.2
@@ -55,10 +54,11 @@ Requires-Dist: pytest; extra == "dev"
 Requires-Dist: pytest-cov; extra == "dev"
 Requires-Dist: coverage[toml]; extra == "dev"
 Requires-Dist: mkdocs>=1.6.0; extra == "dev"
+Requires-Dist: beautifulsoup4<=4.12.3; extra == "dev"
 Requires-Dist: mkdocs-material>=9.5.9; extra == "dev"
 Requires-Dist: mkdocstrings[python]; extra == "dev"
 Requires-Dist: mkdocs-redirects; extra == "dev"
-Requires-Dist: mkdocs-ultralytics-plugin>=0.1.8; extra == "dev"
+Requires-Dist: mkdocs-ultralytics-plugin>=0.1.16; extra == "dev"
 Requires-Dist: mkdocs-macros-plugin>=1.0.5; extra == "dev"
 Provides-Extra: export
 Requires-Dist: onnx>=1.12.0; extra == "export"