ultralytics 8.3.206__tar.gz → 8.3.208__tar.gz

{ultralytics-8.3.206 → ultralytics-8.3.208}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: ultralytics
-Version: 8.3.206
+Version: 8.3.208
 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-8.3.206 → ultralytics-8.3.208}/tests/test_cuda.py

@@ -104,7 +104,7 @@ def test_export_engine_matrix(task, dynamic, int8, half, batch):
     )
     YOLO(file)([SOURCE] * batch, imgsz=64 if dynamic else 32, device=DEVICES[0])  # exported model inference
     Path(file).unlink()  # cleanup
-    Path(file).with_suffix(".cache").unlink() if int8 else None  # cleanup INT8 cache
+    Path(file).with_suffix(".cache").unlink(missing_ok=True) if int8 else None  # cleanup INT8 cache
 
 
 @pytest.mark.skipif(not DEVICES, reason="No CUDA devices available")

{ultralytics-8.3.206 → ultralytics-8.3.208}/tests/test_exports.py

@@ -101,7 +101,7 @@ def test_export_onnx_matrix(task, dynamic, int8, half, batch, simplify, nms):
     [  # generate all combinations except for exclusion cases
         (task, dynamic, int8, half, batch, nms)
         for task, dynamic, int8, half, batch, nms in product(
-            TASKS, [False, True], [False], [False], [1, 2], [True, False]
+            TASKS, [False, True], [False], [False, True], [1, 2], [True, False]
         )
         if not (task == "classify" and nms)
     ],

{ultralytics-8.3.206 → ultralytics-8.3.208}/tests/test_python.py

@@ -408,7 +408,7 @@ def test_utils_init():
 def test_utils_checks():
     """Test various utility checks for filenames, git status, requirements, image sizes, and versions."""
     checks.check_yolov5u_filename("yolov5n.pt")
-    checks.check_requirements()  # check requirements.txt
+    checks.check_requirements("numpy")  # check requirements.txt
     checks.check_imgsz([600, 600], max_dim=1)
     checks.check_imshow(warn=True)
     checks.check_version("ultralytics", "8.0.0")

{ultralytics-8.3.206 → ultralytics-8.3.208}/ultralytics/__init__.py

@@ -1,6 +1,6 @@
 # Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license
 
-__version__ = "8.3.206"
+__version__ = "8.3.208"
 
 import importlib
 import os

{ultralytics-8.3.206 → ultralytics-8.3.208}/ultralytics/engine/exporter.py

@@ -360,8 +360,8 @@ class Exporter:
         if self.args.half and self.args.int8:
             LOGGER.warning("half=True and int8=True are mutually exclusive, setting half=False.")
             self.args.half = False
-        if self.args.half and onnx and self.device.type == "cpu":
-            LOGGER.warning("half=True only compatible with GPU export, i.e. use device=0")
+        if self.args.half and (onnx or jit) and self.device.type == "cpu":
+            LOGGER.warning("half=True only compatible with GPU export, i.e. use device=0, setting half=False.")
             self.args.half = False
         self.imgsz = check_imgsz(self.args.imgsz, stride=model.stride, min_dim=2)  # check image size
         if self.args.optimize:
@@ -463,7 +463,7 @@ class Exporter:
         y = None
         for _ in range(2):  # dry runs
             y = NMSModel(model, self.args)(im) if self.args.nms and not coreml and not imx else model(im)
-        if self.args.half and onnx and self.device.type != "cpu":
+        if self.args.half and (onnx or jit) and self.device.type != "cpu":
             im, model = im.half(), model.half()  # to FP16
 
         # Filter warnings
@@ -510,11 +510,11 @@ class Exporter:
         self.run_callbacks("on_export_start")
         # Exports
         f = [""] * len(fmts)  # exported filenames
-        if jit or ncnn:  # TorchScript
+        if jit:  # TorchScript
             f[0] = self.export_torchscript()
         if engine:  # TensorRT required before ONNX
             f[1] = self.export_engine(dla=dla)
-        if onnx:  # ONNX
+        if onnx or ncnn:  # ONNX
             f[2] = self.export_onnx()
         if xml:  # OpenVINO
             f[3] = self.export_openvino()
@@ -806,7 +806,7 @@ class Exporter:
 
         LOGGER.info(f"\n{prefix} starting export with NCNN {ncnn.__version__}...")
         f = Path(str(self.file).replace(self.file.suffix, f"_ncnn_model{os.sep}"))
-        f_ts = self.file.with_suffix(".torchscript")
+        f_onnx = self.file.with_suffix(".onnx")
 
         name = Path("pnnx.exe" if WINDOWS else "pnnx")  # PNNX filename
         pnnx = name if name.is_file() else (ROOT / name)
@@ -820,10 +820,10 @@ class Exporter:
             try:
                 release, assets = get_github_assets(repo="pnnx/pnnx")
                 asset = [x for x in assets if f"{system}.zip" in x][0]
-                assert isinstance(asset, str), "Unable to retrieve PNNX repo assets"  # i.e. pnnx-20240410-macos.zip
+                assert isinstance(asset, str), "Unable to retrieve PNNX repo assets"  # i.e. pnnx-20250930-macos.zip
                 LOGGER.info(f"{prefix} successfully found latest PNNX asset file {asset}")
             except Exception as e:
-                release = "20240410"
+                release = "20250930"
                 asset = f"pnnx-{release}-{system}.zip"
                 LOGGER.warning(f"{prefix} PNNX GitHub assets not found: {e}, using default {asset}")
             unzip_dir = safe_download(f"https://github.com/pnnx/pnnx/releases/download/{release}/{asset}", delete=True)
@@ -847,7 +847,7 @@ class Exporter:
 
         cmd = [
             str(pnnx),
-            str(f_ts),
+            str(f_onnx),
             *ncnn_args,
             *pnnx_args,
             f"fp16={int(self.args.half)}",
@@ -894,7 +894,7 @@ class Exporter:
             classifier_config = ct.ClassifierConfig(list(self.model.names.values()))
             model = self.model
         elif self.model.task == "detect":
-            model = IOSDetectModel(self.model, self.im) if self.args.nms else self.model
+            model = IOSDetectModel(self.model, self.im, mlprogram=not mlmodel) if self.args.nms else self.model
         else:
             if self.args.nms:
                 LOGGER.warning(f"{prefix} 'nms=True' is only available for Detect models like 'yolo11n.pt'.")
@@ -938,12 +938,7 @@ class Exporter:
                 config = cto.OptimizationConfig(global_config=op_config)
                 ct_model = cto.palettize_weights(ct_model, config=config)
         if self.args.nms and self.model.task == "detect":
-            if mlmodel:
-                weights_dir = None
-            else:
-                ct_model.save(str(f))  # save otherwise weights_dir does not exist
-                weights_dir = str(f / "Data/com.apple.CoreML/weights")
-            ct_model = self._pipeline_coreml(ct_model, weights_dir=weights_dir)
+            ct_model = self._pipeline_coreml(ct_model, weights_dir=None if mlmodel else ct_model.weights_dir)
 
         m = self.metadata  # metadata dict
         ct_model.short_description = m.pop("description")
@@ -1277,7 +1272,8 @@ class Exporter:
         names = self.metadata["names"]
         nx, ny = spec.description.input[0].type.imageType.width, spec.description.input[0].type.imageType.height
         nc = outs[0].type.multiArrayType.shape[-1]
-        assert len(names) == nc, f"{len(names)} names found for nc={nc}"  # check
+        if len(names) != nc:  # Hack fix for MLProgram NMS bug https://github.com/ultralytics/ultralytics/issues/22309
+            names = {**names, **{i: str(i) for i in range(len(names), nc)}}
 
         # Model from spec
         model = ct.models.MLModel(spec, weights_dir=weights_dir)
@@ -1367,18 +1363,20 @@ class Exporter:
 class IOSDetectModel(torch.nn.Module):
     """Wrap an Ultralytics YOLO model for Apple iOS CoreML export."""
 
-    def __init__(self, model, im):
+    def __init__(self, model, im, mlprogram=True):
         """
         Initialize the IOSDetectModel class with a YOLO model and example image.
 
         Args:
             model (torch.nn.Module): The YOLO model to wrap.
             im (torch.Tensor): Example input tensor with shape (B, C, H, W).
+            mlprogram (bool): Whether exporting to MLProgram format to fix NMS bug.
         """
         super().__init__()
         _, _, h, w = im.shape  # batch, channel, height, width
         self.model = model
         self.nc = len(model.names)  # number of classes
+        self.mlprogram = mlprogram
         if w == h:
             self.normalize = 1.0 / w  # scalar
         else:
@@ -1390,7 +1388,11 @@ class IOSDetectModel(torch.nn.Module):
     def forward(self, x):
         """Normalize predictions of object detection model with input size-dependent factors."""
         xywh, cls = self.model(x)[0].transpose(0, 1).split((4, self.nc), 1)
-        return cls, xywh * self.normalize  # confidence (3780, 80), coordinates (3780, 4)
+        if self.mlprogram and self.nc % 80 != 0:  # NMS bug https://github.com/ultralytics/ultralytics/issues/22309
+            pad_length = int(((self.nc + 79) // 80) * 80) - self.nc  # pad class length to multiple of 80
+            cls = torch.nn.functional.pad(cls, (0, pad_length, 0, 0), "constant", 0)
+
+        return cls, xywh * self.normalize
 
 
 class NMSModel(torch.nn.Module):

{ultralytics-8.3.206 → ultralytics-8.3.208}/ultralytics/nn/autobackend.py

@@ -254,7 +254,7 @@ class AutoBackend(nn.Module):
                 session = onnxruntime.InferenceSession(w, providers=providers)
             else:
                 check_requirements(
-                    ["model-compression-toolkit>=2.4.1", "sony-custom-layers[torch]>=0.3.0", "onnxruntime-extensions"]
+                    ("model-compression-toolkit>=2.4.1", "sony-custom-layers[torch]>=0.3.0", "onnxruntime-extensions")
                 )
                 w = next(Path(w).glob("*.onnx"))
                 LOGGER.info(f"Loading {w} for ONNX IMX inference...")

{ultralytics-8.3.206 → ultralytics-8.3.208}/ultralytics/nn/modules/head.py

@@ -896,6 +896,10 @@ class RTDETRDecoder(nn.Module):
     """
 
     export = False  # export mode
+    shapes = []
+    anchors = torch.empty(0)
+    valid_mask = torch.empty(0)
+    dynamic = False
 
     def __init__(
         self,
@@ -1116,10 +1120,12 @@ class RTDETRDecoder(nn.Module):
             enc_scores (torch.Tensor): Encoded scores.
         """
         bs = feats.shape[0]
-        # Prepare input for decoder
-        anchors, valid_mask = self._generate_anchors(shapes, dtype=feats.dtype, device=feats.device)
-        features = self.enc_output(valid_mask * feats)  # bs, h*w, 256
+        if self.dynamic or self.shapes != shapes:
+            self.anchors, self.valid_mask = self._generate_anchors(shapes, dtype=feats.dtype, device=feats.device)
+            self.shapes = shapes
 
+        # Prepare input for decoder
+        features = self.enc_output(self.valid_mask * feats)  # bs, h*w, 256
         enc_outputs_scores = self.enc_score_head(features)  # (bs, h*w, nc)
 
         # Query selection
@@ -1131,7 +1137,7 @@ class RTDETRDecoder(nn.Module):
         # (bs, num_queries, 256)
         top_k_features = features[batch_ind, topk_ind].view(bs, self.num_queries, -1)
         # (bs, num_queries, 4)
-        top_k_anchors = anchors[:, topk_ind].view(bs, self.num_queries, -1)
+        top_k_anchors = self.anchors[:, topk_ind].view(bs, self.num_queries, -1)
 
         # Dynamic anchors + static content
         refer_bbox = self.enc_bbox_head(top_k_features) + top_k_anchors

{ultralytics-8.3.206 → ultralytics-8.3.208}/ultralytics/utils/benchmarks.py

@@ -595,7 +595,7 @@ class ProfileModels:
             mean_time (float): Mean inference time in milliseconds.
             std_time (float): Standard deviation of inference time in milliseconds.
         """
-        check_requirements("onnxruntime")
+        check_requirements([("onnxruntime", "onnxruntime-gpu")])  # either package meets requirements
         import onnxruntime as ort
 
         # Session with either 'TensorrtExecutionProvider', 'CUDAExecutionProvider', 'CPUExecutionProvider'

{ultralytics-8.3.206 → ultralytics-8.3.208}/ultralytics/utils/checks.py

@@ -363,8 +363,9 @@ def check_requirements(requirements=ROOT.parent / "requirements.txt", exclude=()
     Check if installed dependencies meet Ultralytics YOLO models requirements and attempt to auto-update if needed.
 
     Args:
-        requirements (Path | str | list[str] | tuple[str]): Path to a requirements.txt file, a single package
-            requirement as a string, or a list of package requirements as strings.
+        requirements (Path | str | list[str|tuple] | tuple[str]): Path to a requirements.txt file, a single package
+            requirement as a string, a list of package requirements as strings, or a list containing strings and
+            tuples of interchangeable packages.
         exclude (tuple): Tuple of package names to exclude from checking.
         install (bool): If True, attempt to auto-update packages that don't meet requirements.
         cmds (str): Additional commands to pass to the pip install command when auto-updating.
@@ -376,10 +377,13 @@ def check_requirements(requirements=ROOT.parent / "requirements.txt", exclude=()
         >>> check_requirements("path/to/requirements.txt")
 
         Check a single package
-        >>> check_requirements("ultralytics>=8.0.0")
+        >>> check_requirements("ultralytics>=8.3.200", cmds="--index-url https://download.pytorch.org/whl/cpu")
 
         Check multiple packages
-        >>> check_requirements(["numpy", "ultralytics>=8.0.0"])
+        >>> check_requirements(["numpy", "ultralytics"])
+
+        Check with interchangeable packages
+        >>> check_requirements([("onnxruntime", "onnxruntime-gpu"), "numpy"])
     """
     prefix = colorstr("red", "bold", "requirements:")
     if isinstance(requirements, Path):  # requirements.txt file
@@ -391,13 +395,22 @@ def check_requirements(requirements=ROOT.parent / "requirements.txt", exclude=()
 
     pkgs = []
     for r in requirements:
-        r_stripped = r.rpartition("/")[-1].replace(".git", "")  # replace git+https://org/repo.git -> 'repo'
-        match = re.match(r"([a-zA-Z0-9-_]+)([<>!=~]+.*)?", r_stripped)
-        name, required = match[1], match[2].strip() if match[2] else ""
-        try:
-            assert check_version(metadata.version(name), required)  # exception if requirements not met
-        except (AssertionError, metadata.PackageNotFoundError):
-            pkgs.append(r)
+        candidates = r if isinstance(r, (list, tuple)) else [r]
+        satisfied = False
+
+        for candidate in candidates:
+            r_stripped = candidate.rpartition("/")[-1].replace(".git", "")  # replace git+https://org/repo.git -> 'repo'
+            match = re.match(r"([a-zA-Z0-9-_]+)([<>!=~]+.*)?", r_stripped)
+            name, required = match[1], match[2].strip() if match[2] else ""
+            try:
+                if check_version(metadata.version(name), required):
+                    satisfied = True
+                    break
+            except (AssertionError, metadata.PackageNotFoundError):
+                continue
+
+        if not satisfied:
+            pkgs.append(candidates[0])
 
     @Retry(times=2, delay=1)
     def attempt_install(packages, commands, use_uv):

{ultralytics-8.3.206 → ultralytics-8.3.208}/ultralytics/utils/downloads.py

@@ -374,10 +374,10 @@ def safe_download(
                 raise  # Re-raise immediately - no point retrying if insufficient disk space
             except Exception as e:
                 if i == 0 and not is_online():
-                    raise ConnectionError(emojis(f"❌  Download failure for {uri}. Environment is not online.")) from e
+                    raise ConnectionError(emojis(f"❌  Download failure for {uri}. Environment may be offline.")) from e
                 elif i >= retry:
-                    raise ConnectionError(emojis(f"❌  Download failure for {uri}. Retry limit reached.")) from e
-                LOGGER.warning(f"Download failure, retrying {i + 1}/{retry} {uri}...")
+                    raise ConnectionError(emojis(f"❌  Download failure for {uri}. Retry limit reached. {e}")) from e
+                LOGGER.warning(f"Download failure, retrying {i + 1}/{retry} {uri}... {e}")
 
     if unzip and f.exists() and f.suffix in {"", ".zip", ".tar", ".gz"}:
         from zipfile import is_zipfile

{ultralytics-8.3.206 → ultralytics-8.3.208}/ultralytics.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: ultralytics
-Version: 8.3.206
+Version: 8.3.208
 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>