ultralytics 8.3.140__tar.gz → 8.3.141__tar.gz

{ultralytics-8.3.140/ultralytics.egg-info → ultralytics-8.3.141}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: ultralytics
-Version: 8.3.140
+Version: 8.3.141
 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-8.3.140 → ultralytics-8.3.141}/tests/test_cuda.py

@@ -22,7 +22,7 @@ if CUDA_IS_AVAILABLE:
     else:
         gpu_info = GPUInfo()
         gpu_info.print_status()
-        idle_gpus = gpu_info.select_idle_gpu(count=2, min_memory_mb=2048)
+        idle_gpus = gpu_info.select_idle_gpu(count=2, min_memory_fraction=0.2)
         if idle_gpus:
             DEVICES = idle_gpus
 

{ultralytics-8.3.140 → ultralytics-8.3.141}/tests/test_solutions.py

@@ -205,12 +205,12 @@ def test_solution(name, solution_class, needs_frame_count, video, kwargs):
     )
 
 
-@pytest.mark.slow
 @pytest.mark.skipif(checks.IS_PYTHON_3_8, reason="Disabled due to unsupported CLIP dependencies.")
 @pytest.mark.skipif(IS_RASPBERRYPI, reason="Disabled due to slow performance on Raspberry Pi.")
 def test_similarity_search():
     """Test similarity search solution."""
-    searcher = solutions.VisualAISearch()
+    safe_download(f"{ASSETS_URL}/4-imgs-similaritysearch.zip", dir=TMP)  # 4 dog images for testing in a zip file.
+    searcher = solutions.VisualAISearch(data=str(TMP / "4-imgs-similaritysearch"))
     _ = searcher("a dog sitting on a bench")  # Returns the results in format "- img name | similarity score"
 
 
@@ -297,6 +297,7 @@ def test_streamlit_handle_video_upload_creates_file():
     os.remove("ultralytics.mp4")
 
 
+@pytest.mark.skipif(checks.IS_PYTHON_3_8, reason="Disabled due to unsupported CLIP dependencies.")
 @pytest.mark.skipif(IS_RASPBERRYPI, reason="Disabled due to slow performance on Raspberry Pi.")
 def test_similarity_search_app_init():
     """Test SearchApp initializes with required attributes."""

{ultralytics-8.3.140 → ultralytics-8.3.141}/ultralytics/__init__.py

@@ -1,6 +1,6 @@
 # Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license
 
-__version__ = "8.3.140"
+__version__ = "8.3.141"
 
 import os
 

{ultralytics-8.3.140 → ultralytics-8.3.141}/ultralytics/cfg/datasets/open-images-v7.yaml

@@ -622,21 +622,18 @@ names:
 download: |
   import warnings
 
-  from ultralytics.utils import LOGGER, SETTINGS, Path, get_ubuntu_version, is_ubuntu
-  from ultralytics.utils.checks import check_requirements, check_version
+  from ultralytics.utils import LOGGER, SETTINGS, Path
+  from ultralytics.utils.checks import check_requirements
 
   check_requirements("fiftyone")
-  if is_ubuntu() and check_version(get_ubuntu_version(), ">=22.04"):
-      # Ubuntu>=22.04 patch https://github.com/voxel51/fiftyone/issues/2961#issuecomment-1666519347
-      check_requirements("fiftyone-db-ubuntu2204")
-    
+
   import fiftyone as fo
   import fiftyone.zoo as foz
 
   name = "open-images-v7"
   fo.config.dataset_zoo_dir = Path(SETTINGS["datasets_dir"]) / "fiftyone" / name
   fraction = 1.0  # fraction of full dataset to use
-  LOGGER.warning("WARNING ⚠️ Open Images V7 dataset requires at least **561 GB of free space. Starting download...")
+  LOGGER.warning("Open Images V7 dataset requires at least **561 GB of free space. Starting download...")
   for split in "train", "validation":  # 1743042 train, 41620 val images
       train = split == "train"
 

{ultralytics-8.3.140 → ultralytics-8.3.141}/ultralytics/engine/model.py

@@ -79,7 +79,7 @@ class Model(torch.nn.Module):
 
     def __init__(
         self,
-        model: Union[str, Path] = "yolo11n.pt",
+        model: Union[str, Path, "Model"] = "yolo11n.pt",
         task: str = None,
         verbose: bool = False,
     ) -> None:
@@ -92,8 +92,8 @@ class Model(torch.nn.Module):
         prediction, or export.
 
         Args:
-            model (str | Path): Path or name of the model to load or create. Can be a local file path, a
-                model name from Ultralytics HUB, or a Triton Server model.
+            model (str | Path | Model): Path or name of the model to load or create. Can be a local file path, a
+                model name from Ultralytics HUB, a Triton Server model, or an already initialized Model instance.
             task (str | None): The task type associated with the YOLO model, specifying its application domain.
             verbose (bool): If True, enables verbose output during the model's initialization and subsequent
                 operations.
@@ -108,6 +108,9 @@ class Model(torch.nn.Module):
             >>> model = Model("path/to/model.yaml", task="detect")
             >>> model = Model("hub_model", verbose=True)
         """
+        if isinstance(model, Model):
+            self.__dict__ = model.__dict__  # accepts an already initialized Model
+            return
         super().__init__()
         self.callbacks = callbacks.get_default_callbacks()
         self.predictor = None  # reuse predictor

{ultralytics-8.3.140 → ultralytics-8.3.141}/ultralytics/models/yolo/classify/predict.py

@@ -57,7 +57,7 @@ class ClassificationPredictor(BasePredictor):
         super().setup_source(source)
         updated = (
             self.model.model.transforms.transforms[0].size != max(self.imgsz)
-            if hasattr(self.model.model, "transforms")
+            if hasattr(self.model.model, "transforms") and hasattr(self.model.model.transforms.transforms[0], "size")
             else True
         )
         self.transforms = self.model.model.transforms if not updated else classify_transforms(self.imgsz)

{ultralytics-8.3.140 → ultralytics-8.3.141}/ultralytics/models/yolo/model.py

@@ -39,7 +39,7 @@ class YOLO(Model):
             >>> model = YOLO("yolo11n.pt")  # load a pretrained YOLOv11n detection model
             >>> model = YOLO("yolo11n-seg.pt")  # load a pretrained YOLO11n segmentation model
         """
-        path = Path(model)
+        path = Path(model if isinstance(model, (str, Path)) else "")
         if "-world" in path.stem and path.suffix in {".pt", ".yaml", ".yml"}:  # if YOLOWorld PyTorch model
             new_instance = YOLOWorld(path, verbose=verbose)
             self.__class__ = type(new_instance)
@@ -51,6 +51,12 @@ class YOLO(Model):
         else:
             # Continue with default YOLO initialization
             super().__init__(model=model, task=task, verbose=verbose)
+            if hasattr(self.model, "model") and "RTDETR" in self.model.model[-1]._get_name():  # if RTDETR head
+                from ultralytics import RTDETR
+
+                new_instance = RTDETR(self)
+                self.__class__ = type(new_instance)
+                self.__dict__ = new_instance.__dict__
 
     @property
     def task_map(self):

{ultralytics-8.3.140 → ultralytics-8.3.141}/ultralytics/utils/autodevice.py

@@ -116,13 +116,13 @@ class GPUInfo:
 
         LOGGER.info(f"{'-' * len(hdr)}\n")
 
-    def select_idle_gpu(self, count=1, min_memory_mb=0):
+    def select_idle_gpu(self, count=1, min_memory_fraction=0):
         """
         Selects the 'count' most idle GPUs based on utilization and free memory.
 
         Args:
             count (int): The number of idle GPUs to select. Defaults to 1.
-            min_memory_mb (int): Minimum free memory required (MiB). Defaults to 0.
+            min_memory_fraction (float): Minimum free memory required (fraction). Defaults to 0.
 
         Returns:
             (list[int]): Indices of the selected GPUs, sorted by idleness.
@@ -131,7 +131,8 @@ class GPUInfo:
              Returns fewer than 'count' if not enough qualify or exist.
              Returns basic CUDA indices if NVML fails. Empty list if no GPUs found.
         """
-        LOGGER.info(f"Searching for {count} idle GPUs with >= {min_memory_mb} MiB free memory...")
+        assert min_memory_fraction <= 1.0, f"min_memory_fraction must be <= 1.0, got {min_memory_fraction}"
+        LOGGER.info(f"Searching for {count} idle GPUs with >= {min_memory_fraction * 100:.1f}% free memory...")
 
         if count <= 0:
             return []
@@ -145,7 +146,8 @@ class GPUInfo:
         eligible_gpus = [
             gpu
             for gpu in self.gpu_stats
-            if gpu.get("memory_free", -1) >= min_memory_mb and gpu.get("utilization", -1) != -1
+            if gpu.get("memory_free", 0) / gpu.get("memory_total", 1) >= min_memory_fraction
+            and gpu.get("utilization", -1) != -1
         ]
         eligible_gpus.sort(key=lambda x: (x.get("utilization", 101), -x.get("memory_free", 0)))
 
@@ -155,19 +157,19 @@ class GPUInfo:
         if selected:
             LOGGER.info(f"Selected idle CUDA devices {selected}")
         else:
-            LOGGER.warning(f"No GPUs met criteria (Util != -1, Free Mem >= {min_memory_mb} MiB).")
+            LOGGER.warning(f"No GPUs met criteria (Util != -1, Free Mem >= {min_memory_fraction * 100:.1f}%).")
 
         return selected
 
 
 if __name__ == "__main__":
-    required_free_mem = 2048  # Require 2GB free VRAM
+    required_free_mem_fraction = 0.2  # Require 20% free VRAM
     num_gpus_to_select = 1
 
     gpu_info = GPUInfo()
     gpu_info.print_status()
 
-    selected = gpu_info.select_idle_gpu(count=num_gpus_to_select, min_memory_mb=required_free_mem)
+    selected = gpu_info.select_idle_gpu(count=num_gpus_to_select, min_memory_fraction=required_free_mem_fraction)
     if selected:
         print(f"\n==> Using selected GPU indices: {selected}")
         devices = [f"cuda:{idx}" for idx in selected]

{ultralytics-8.3.140 → ultralytics-8.3.141}/ultralytics/utils/loss.py

@@ -613,8 +613,7 @@ class v8ClassificationLoss:
         """Compute the classification loss between predictions and true labels."""
         preds = preds[1] if isinstance(preds, (list, tuple)) else preds
         loss = F.cross_entropy(preds, batch["cls"], reduction="mean")
-        loss_items = loss.detach()
-        return loss, loss_items
+        return loss, loss.detach()
 
 
 class v8OBBLoss(v8DetectionLoss):

{ultralytics-8.3.140 → ultralytics-8.3.141}/ultralytics/utils/tal.py

@@ -21,7 +21,6 @@ class TaskAlignedAssigner(nn.Module):
     Attributes:
         topk (int): The number of top candidates to consider.
         num_classes (int): The number of object classes.
-        bg_idx (int): Background class index.
         alpha (float): The alpha parameter for the classification component of the task-aligned metric.
         beta (float): The beta parameter for the localization component of the task-aligned metric.
         eps (float): A small value to prevent division by zero.
@@ -32,7 +31,6 @@ class TaskAlignedAssigner(nn.Module):
         super().__init__()
         self.topk = topk
         self.num_classes = num_classes
-        self.bg_idx = num_classes
         self.alpha = alpha
         self.beta = beta
         self.eps = eps
@@ -66,7 +64,7 @@ class TaskAlignedAssigner(nn.Module):
 
         if self.n_max_boxes == 0:
             return (
-                torch.full_like(pd_scores[..., 0], self.bg_idx),
+                torch.full_like(pd_scores[..., 0], self.num_classes),
                 torch.zeros_like(pd_bboxes),
                 torch.zeros_like(pd_scores),
                 torch.zeros_like(pd_scores[..., 0]),
@@ -193,7 +191,7 @@ class TaskAlignedAssigner(nn.Module):
         """
         return bbox_iou(gt_bboxes, pd_bboxes, xywh=False, CIoU=True).squeeze(-1).clamp_(0)
 
-    def select_topk_candidates(self, metrics, largest=True, topk_mask=None):
+    def select_topk_candidates(self, metrics, topk_mask=None):
         """
         Select the top-k candidates based on the given metrics.
 
@@ -201,7 +199,6 @@ class TaskAlignedAssigner(nn.Module):
             metrics (torch.Tensor): A tensor of shape (b, max_num_obj, h*w), where b is the batch size,
                               max_num_obj is the maximum number of objects, and h*w represents the
                               total number of anchor points.
-            largest (bool): If True, select the largest values; otherwise, select the smallest values.
             topk_mask (torch.Tensor): An optional boolean tensor of shape (b, max_num_obj, topk), where
                                 topk is the number of top candidates to consider. If not provided,
                                 the top-k values are automatically computed based on the given metrics.
@@ -210,7 +207,7 @@ class TaskAlignedAssigner(nn.Module):
             (torch.Tensor): A tensor of shape (b, max_num_obj, h*w) containing the selected top-k candidates.
         """
         # (b, max_num_obj, topk)
-        topk_metrics, topk_idxs = torch.topk(metrics, self.topk, dim=-1, largest=largest)
+        topk_metrics, topk_idxs = torch.topk(metrics, self.topk, dim=-1, largest=True)
         if topk_mask is None:
             topk_mask = (topk_metrics.max(-1, keepdim=True)[0] > self.eps).expand_as(topk_idxs)
         # (b, max_num_obj, topk)

{ultralytics-8.3.140 → ultralytics-8.3.141}/ultralytics/utils/torch_utils.py

@@ -174,7 +174,7 @@ def select_device(device="", batch=0, newline=False, verbose=True):
 
         # Replace each -1 with a selected GPU or remove it
         parts = device.split(",")
-        selected = GPUInfo().select_idle_gpu(count=parts.count("-1"), min_memory_mb=2048)
+        selected = GPUInfo().select_idle_gpu(count=parts.count("-1"), min_memory_fraction=0.2)
         for i in range(len(parts)):
             if parts[i] == "-1":
                 parts[i] = str(selected.pop(0)) if selected else ""

{ultralytics-8.3.140 → ultralytics-8.3.141/ultralytics.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: ultralytics
-Version: 8.3.140
+Version: 8.3.141
 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>