ultralytics 8.3.1__tar.gz → 8.3.2__tar.gz

{ultralytics-8.3.1 → ultralytics-8.3.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: ultralytics
-Version: 8.3.1
+Version: 8.3.2
 Summary: Ultralytics YOLO for SOTA object detection, multi-object tracking, instance segmentation, pose estimation and image classification.
 Author: Ayush Chaurasia
 Author-email: Glenn Jocher <glenn.jocher@ultralytics.com>, Jing Qiu <jing.qiu@ultralytics.com>

{ultralytics-8.3.1 → ultralytics-8.3.2}/tests/test_cuda.py

@@ -10,6 +10,7 @@ from tests import CUDA_DEVICE_COUNT, CUDA_IS_AVAILABLE, MODEL, SOURCE
 from ultralytics import YOLO
 from ultralytics.cfg import TASK2DATA, TASK2MODEL, TASKS
 from ultralytics.utils import ASSETS, WEIGHTS_DIR
+from ultralytics.utils.checks import check_amp
 
 
 def test_checks():
@@ -18,6 +19,13 @@ def test_checks():
     assert torch.cuda.device_count() == CUDA_DEVICE_COUNT
 
 
+@pytest.mark.skipif(not CUDA_IS_AVAILABLE, reason="CUDA is not available")
+def test_amp():
+    """Test AMP training checks."""
+    model = YOLO("yolo11n.pt").model.cuda()
+    assert check_amp(model)
+
+
 @pytest.mark.slow
 @pytest.mark.skipif(True, reason="CUDA export tests disabled pending additional Ultralytics GPU server availability")
 @pytest.mark.skipif(not CUDA_IS_AVAILABLE, reason="CUDA is not available")

{ultralytics-8.3.1 → ultralytics-8.3.2}/ultralytics/__init__.py

@@ -1,6 +1,6 @@
 # Ultralytics YOLO 🚀, AGPL-3.0 license
 
-__version__ = "8.3.1"
+__version__ = "8.3.2"
 
 import os
 

{ultralytics-8.3.1 → ultralytics-8.3.2}/ultralytics/engine/model.py

@@ -377,7 +377,7 @@ class Model(nn.Module):
         self.model.load(weights)
         return self
 
-    def save(self, filename: Union[str, Path] = "saved_model.pt", use_dill=True) -> None:
+    def save(self, filename: Union[str, Path] = "saved_model.pt") -> None:
         """
         Saves the current model state to a file.
 
@@ -386,7 +386,6 @@ class Model(nn.Module):
 
         Args:
             filename (Union[str, Path]): The name of the file to save the model to.
-            use_dill (bool): Whether to try using dill for serialization if available.
 
         Raises:
             AssertionError: If the model is not a PyTorch model.
@@ -408,7 +407,7 @@ class Model(nn.Module):
             "license": "AGPL-3.0 License (https://ultralytics.com/license)",
             "docs": "https://docs.ultralytics.com",
         }
-        torch.save({**self.ckpt, **updates}, filename, use_dill=use_dill)
+        torch.save({**self.ckpt, **updates}, filename)
 
     def info(self, detailed: bool = False, verbose: bool = True):
         """

{ultralytics-8.3.1 → ultralytics-8.3.2}/ultralytics/models/sam/build.py

@@ -210,8 +210,6 @@ def _build_sam(
             state_dict = torch.load(f)
         sam.load_state_dict(state_dict)
     sam.eval()
-    # sam.load_state_dict(torch.load(checkpoint), strict=True)
-    # sam.eval()
     return sam
 
 

{ultralytics-8.3.1 → ultralytics-8.3.2}/ultralytics/utils/__init__.py

@@ -111,6 +111,7 @@ torch.set_printoptions(linewidth=320, precision=4, profile="default")
 np.set_printoptions(linewidth=320, formatter={"float_kind": "{:11.5g}".format})  # format short g, %precision=5
 cv2.setNumThreads(0)  # prevent OpenCV from multithreading (incompatible with PyTorch DataLoader)
 os.environ["NUMEXPR_MAX_THREADS"] = str(NUM_THREADS)  # NumExpr max threads
+os.environ["CUBLAS_WORKSPACE_CONFIG"] = ":4096:8"  # for deterministic training to avoid CUDA warning
 os.environ["TF_CPP_MIN_LOG_LEVEL"] = "3"  # suppress verbose TF compiler warnings in Colab
 os.environ["TORCH_CPP_LOG_LEVEL"] = "ERROR"  # suppress "NNPACK.cpp could not initialize NNPACK" warnings
 os.environ["KINETO_LOG_LEVEL"] = "5"  # suppress verbose PyTorch profiler output when computing FLOPs

{ultralytics-8.3.1 → ultralytics-8.3.2}/ultralytics/utils/checks.py

@@ -657,9 +657,10 @@ def check_amp(model):
     def amp_allclose(m, im):
         """All close FP32 vs AMP results."""
         batch = [im] * 8
-        a = m(batch, imgsz=128, device=device, verbose=False)[0].boxes.data  # FP32 inference
+        imgsz = max(256, int(model.stride.max() * 4))  # max stride P5-32 and P6-64
+        a = m(batch, imgsz=imgsz, device=device, verbose=False)[0].boxes.data  # FP32 inference
         with autocast(enabled=True):
-            b = m(batch, imgsz=128, device=device, verbose=False)[0].boxes.data  # AMP inference
+            b = m(batch, imgsz=imgsz, device=device, verbose=False)[0].boxes.data  # AMP inference
         del m
         return a.shape == b.shape and torch.allclose(a, b.float(), atol=0.5)  # close to 0.5 absolute tolerance
 

{ultralytics-8.3.1 → ultralytics-8.3.2}/ultralytics/utils/files.py

@@ -219,4 +219,4 @@ def update_models(model_names=("yolov8n.pt",), source_dir=Path("."), update_name
 
         # Save model using model.save()
         print(f"Re-saving {model_name} model to {save_path}")
-        model.save(save_path, use_dill=False)
+        model.save(save_path)

{ultralytics-8.3.1 → ultralytics-8.3.2}/ultralytics/utils/patches.py

@@ -86,25 +86,15 @@ def torch_load(*args, **kwargs):
     return _torch_load(*args, **kwargs)
 
 
-def torch_save(*args, use_dill=True, **kwargs):
+def torch_save(*args, **kwargs):
     """
     Optionally use dill to serialize lambda functions where pickle does not, adding robustness with 3 retries and
     exponential standoff in case of save failure.
 
     Args:
         *args (tuple): Positional arguments to pass to torch.save.
-        use_dill (bool): Whether to try using dill for serialization if available. Defaults to True.
         **kwargs (Any): Keyword arguments to pass to torch.save.
     """
-    try:
-        assert use_dill
-        import dill as pickle
-    except (AssertionError, ImportError):
-        import pickle
-
-    if "pickle_module" not in kwargs:
-        kwargs["pickle_module"] = pickle
-
     for i in range(4):  # 3 retries
         try:
             return _torch_save(*args, **kwargs)

{ultralytics-8.3.1 → ultralytics-8.3.2}/ultralytics/utils/plotting.py

@@ -13,8 +13,8 @@ import torch
 from PIL import Image, ImageDraw, ImageFont
 from PIL import __version__ as pil_version
 
-from ultralytics.utils import LOGGER, TryExcept, ops, plt_settings, threaded
-from ultralytics.utils.checks import check_font, check_version, is_ascii
+from ultralytics.utils import IS_JUPYTER, LOGGER, TryExcept, ops, plt_settings, threaded
+from ultralytics.utils.checks import check_font, check_requirements, check_version, is_ascii
 from ultralytics.utils.files import increment_path
 
 
@@ -524,7 +524,18 @@ class Annotator:
 
     def show(self, title=None):
         """Show the annotated image."""
-        Image.fromarray(np.asarray(self.im)[..., ::-1]).show(title)
+        im = Image.fromarray(np.asarray(self.im)[..., ::-1])  # Convert numpy array to PIL Image with RGB to BGR
+        if IS_JUPYTER:
+            check_requirements("ipython")
+            try:
+                from IPython.display import display
+
+                display(im)
+            except ImportError as e:
+                LOGGER.warning(f"Unable to display image in Jupyter notebooks: {e}")
+        else:
+            # Convert numpy array to PIL Image and show
+            im.show(title=title)
 
     def save(self, filename="image.jpg"):
         """Save the annotated image to 'filename'."""

{ultralytics-8.3.1 → ultralytics-8.3.2}/ultralytics/utils/torch_utils.py

@@ -595,7 +595,7 @@ def strip_optimizer(f: Union[str, Path] = "best.pt", s: str = "", updates: dict
 
     # Save
     combined = {**metadata, **x, **(updates or {})}
-    torch.save(combined, s or f, use_dill=False)  # combine dicts (prefer to the right)
+    torch.save(combined, s or f)  # combine dicts (prefer to the right)
     mb = os.path.getsize(s or f) / 1e6  # file size
     LOGGER.info(f"Optimizer stripped from {f},{f' saved as {s},' if s else ''} {mb:.1f}MB")
     return combined

{ultralytics-8.3.1 → ultralytics-8.3.2}/ultralytics.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: ultralytics
-Version: 8.3.1
+Version: 8.3.2
 Summary: Ultralytics YOLO for SOTA object detection, multi-object tracking, instance segmentation, pose estimation and image classification.
 Author: Ayush Chaurasia
 Author-email: Glenn Jocher <glenn.jocher@ultralytics.com>, Jing Qiu <jing.qiu@ultralytics.com>