dgenerate-ultralytics-headless 8.3.178__py3-none-any.whl → 8.3.180__py3-none-any.whl

{dgenerate_ultralytics_headless-8.3.178.dist-info → dgenerate_ultralytics_headless-8.3.180.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: dgenerate-ultralytics-headless
-Version: 8.3.178
+Version: 8.3.180
 Summary: Automatically built Ultralytics package with python-opencv-headless dependency instead of python-opencv
 Author-email: Glenn Jocher <glenn.jocher@ultralytics.com>, Jing Qiu <jing.qiu@ultralytics.com>
 Maintainer-email: Ultralytics <hello@ultralytics.com>
@@ -57,13 +57,13 @@ Requires-Dist: mkdocs-material>=9.5.9; extra == "dev"
 Requires-Dist: mkdocstrings[python]; extra == "dev"
 Requires-Dist: mkdocs-ultralytics-plugin>=0.1.26; extra == "dev"
 Requires-Dist: mkdocs-macros-plugin>=1.0.5; extra == "dev"
-Requires-Dist: click==8.2.1; extra == "dev"
 Provides-Extra: export
+Requires-Dist: numpy<2.0.0; extra == "export"
 Requires-Dist: onnx<1.18.0,>=1.12.0; extra == "export"
 Requires-Dist: coremltools>=8.0; (platform_system != "Windows" and python_version <= "3.13") and extra == "export"
 Requires-Dist: scikit-learn>=1.3.2; (platform_system != "Windows" and python_version <= "3.13") and extra == "export"
 Requires-Dist: openvino>=2024.0.0; extra == "export"
-Requires-Dist: tensorflow>=2.0.0; extra == "export"
+Requires-Dist: tensorflow<=2.19.0,>=2.0.0; extra == "export"
 Requires-Dist: tensorflowjs>=2.0.0; extra == "export"
 Requires-Dist: tensorstore>=0.1.63; (platform_machine == "aarch64" and python_version >= "3.9") and extra == "export"
 Requires-Dist: h5py!=3.11.0; platform_machine == "aarch64" and extra == "export"

{dgenerate_ultralytics_headless-8.3.178.dist-info → dgenerate_ultralytics_headless-8.3.180.dist-info}/RECORD

@@ -1,4 +1,4 @@
-dgenerate_ultralytics_headless-8.3.178.dist-info/licenses/LICENSE,sha256=DZak_2itbUtvHzD3E7GNUYSRK6jdOJ-GqncQ2weavLA,34523
+dgenerate_ultralytics_headless-8.3.180.dist-info/licenses/LICENSE,sha256=DZak_2itbUtvHzD3E7GNUYSRK6jdOJ-GqncQ2weavLA,34523
 tests/__init__.py,sha256=b4KP5_q-2IO8Br8YHOSLYnn7IwZS81l_vfEF2YPa2lM,894
 tests/conftest.py,sha256=LXtQJcFNWPGuzauTGkiXgsvVC3llJKfg22WcmhRzuQc,2593
 tests/test_cli.py,sha256=EMf5gTAopOnIz8VvzaM-Qb044o7D0flnUHYQ-2ffOM4,5670
@@ -8,10 +8,10 @@ tests/test_exports.py,sha256=CY-4xVZlVM16vdyIC0mSR3Ix59aiZm1qjFGIhSNmB20,11007
 tests/test_integrations.py,sha256=kl_AKmE_Qs1GB0_91iVwbzNxofm_hFTt0zzU6JF-pg4,6323
 tests/test_python.py,sha256=-qvdeg-hEcKU5mWSDEU24iFZ-i8FAwQRznSXpkp6WQ4,27928
 tests/test_solutions.py,sha256=tuf6n_fsI8KvSdJrnc-cqP2qYdiYqCWuVrx0z9dOz3Q,13213
-ultralytics/__init__.py,sha256=jfnesIVlI3MP1Fu1VhXuCSl2_tHCed9GUcoR23EvLlI,730
+ultralytics/__init__.py,sha256=fHQo0GhHdl2c_XkrtYaXY22EnCd1IZeIe5C59ZLWimc,730
 ultralytics/assets/bus.jpg,sha256=wCAZxJecGR63Od3ZRERe9Aja1Weayrb9Ug751DS_vGM,137419
 ultralytics/assets/zidane.jpg,sha256=Ftc4aeMmen1O0A3o6GCDO9FlfBslLpTAw0gnetx7bts,50427
-ultralytics/cfg/__init__.py,sha256=VIpPHImhjb0XLJquGZrG_LBGZchtOtBSXR7HYTYV2GU,39602
+ultralytics/cfg/__init__.py,sha256=Uj1br3-NVFvP6VY5CL4PK63mAQAom93XFC5cqSbM6t4,39887
 ultralytics/cfg/default.yaml,sha256=1SspGAK_K_DT7DBfEScJh4jsJUTOxahehZYj92xmj7o,8347
 ultralytics/cfg/datasets/Argoverse.yaml,sha256=4SGaJio9JFUkrscHJTPnH_QSbYm48Wbk8EFwl39zntc,3262
 ultralytics/cfg/datasets/DOTAv1.5.yaml,sha256=VZ_KKFX0H2YvlFVJ8JHcLWYBZ2xiQ6Z-ROSTiKWpS7c,1211
@@ -21,8 +21,8 @@ ultralytics/cfg/datasets/HomeObjects-3K.yaml,sha256=xEtSqEad-rtfGuIrERjjhdISggmP
 ultralytics/cfg/datasets/ImageNet.yaml,sha256=GvDWypLVG_H3H67Ai8IC1pvK6fwcTtF5FRhzO1OXXDU,42530
 ultralytics/cfg/datasets/Objects365.yaml,sha256=vLzbT3xgpLR-bHhrHOiYyzYvDIniRdevgSyPetm8QHk,9354
 ultralytics/cfg/datasets/SKU-110K.yaml,sha256=a52le1-JQ2YH6b1WLMUxVz7RkZ36YsmXgWyw0z3q9nQ,2542
-ultralytics/cfg/datasets/VOC.yaml,sha256=GfJkYxN6uAiBTHOsR57L0UDi5NE9vH59A15EROrp0DU,3785
-ultralytics/cfg/datasets/VisDrone.yaml,sha256=NujUSnR6gpXYdcvgg9nxmSZjPjcC9MdZ_YzMipvnuK8,3615
+ultralytics/cfg/datasets/VOC.yaml,sha256=o09FWAAsr1MH3ftBJ_n-4Tmc3zxnVJL1HqlqKRUYVTQ,3774
+ultralytics/cfg/datasets/VisDrone.yaml,sha256=dYAewe84CrGmxAA_z6UnZUAd7peaw5l3ARDcssojADk,3604
 ultralytics/cfg/datasets/african-wildlife.yaml,sha256=SuloMp9WAZBigGC8az-VLACsFhTM76_O29yhTvUqdnU,915
 ultralytics/cfg/datasets/brain-tumor.yaml,sha256=qrxPO_t9wxbn2kHFwP3vGTzSWj2ELTLelUwYL3_b6nc,800
 ultralytics/cfg/datasets/carparts-seg.yaml,sha256=A4e9hM1unTY2jjZIXGiKSarF6R-Ad9R99t57OgRJ37w,1253
@@ -151,7 +151,7 @@ ultralytics/models/sam/__init__.py,sha256=iR7B06rAEni21eptg8n4rLOP0Z_qV9y9PL-L93
 ultralytics/models/sam/amg.py,sha256=IpcuIfC5KBRiF4sdrsPl1ecWEJy75axo1yG23r5BFsw,11783
 ultralytics/models/sam/build.py,sha256=J6n-_QOYLa63jldEZmhRe9D3Is_AJE8xyZLUjzfRyTY,12629
 ultralytics/models/sam/model.py,sha256=j1TwsLmtxhiXyceU31VPzGVkjRXGylphKrdPSzUJRJc,7231
-ultralytics/models/sam/predict.py,sha256=2dg6L8X_I4RqTHAeH8w3m2ojFczkplx1Wu_ytwzAAgQ,82979
+ultralytics/models/sam/predict.py,sha256=VBClks5F74BX0uZWpMfUjaXM_H222BYJ09lRYF6A2u8,86184
 ultralytics/models/sam/modules/__init__.py,sha256=lm6MckFYCPTbqIoX7w0s_daxdjNeBeKW6DXppv1-QUM,70
 ultralytics/models/sam/modules/blocks.py,sha256=n8oe9sx91_RktsF2_2UYNKH7qk8bFXuJtEaIEpQQ3ws,46059
 ultralytics/models/sam/modules/decoders.py,sha256=-1fhBO47hA-3CzkU-PzkCK4Nsi_VJ_CH6Q9SMjydN4I,25609
@@ -165,7 +165,7 @@ ultralytics/models/utils/__init__.py,sha256=lm6MckFYCPTbqIoX7w0s_daxdjNeBeKW6DXp
 ultralytics/models/utils/loss.py,sha256=E-61TfLPc04IdeL6IlFDityDoPju-ov0ouWV_cNY4Kg,21254
 ultralytics/models/utils/ops.py,sha256=Pr77n8XW25SUEx4X3bBvXcVIbRdJPoaXJuG0KWWawRQ,15253
 ultralytics/models/yolo/__init__.py,sha256=or0j5xvcM0usMlsFTYhNAOcQUri7reD0cD9JR5b7zDk,307
-ultralytics/models/yolo/model.py,sha256=e66CIsSLHbEeGlkEQ1r6WwVDKAoR2nc0-UoGA94z-eM,18544
+ultralytics/models/yolo/model.py,sha256=96PDREUJwDiPb3w4lp2HCesc3c3y1WGyLttOUhUYPxk,18715
 ultralytics/models/yolo/classify/__init__.py,sha256=9--HVaNOfI1K7rn_rRqclL8FUAnpfeBrRqEQIaQw2xM,383
 ultralytics/models/yolo/classify/predict.py,sha256=FqAC2YXe25bRwedMZhF3Lw0waoY-a60xMKELhxApP9I,4149
 ultralytics/models/yolo/classify/train.py,sha256=V-hevc6X7xemnpyru84OfTRA77eNnkVSMEz16_OUvo4,10244
@@ -239,9 +239,9 @@ ultralytics/utils/__init__.py,sha256=KyczXn2SLR0g8xkCVpfNaPkhDCQ4A8vLt99teYu7mss
 ultralytics/utils/autobatch.py,sha256=33m8YgggLIhltDqMXZ5OE-FGs2QiHrl2-LfgY1mI4cw,5119
 ultralytics/utils/autodevice.py,sha256=AvgXFt8c1Cg4icKh0Hbhhz8UmVQ2Wjyfdfkeb2C8zck,8855
 ultralytics/utils/benchmarks.py,sha256=btsi_B0mfLPfhE8GrsBpi79vl7SRam0YYngNFAsY8Ak,31035
-ultralytics/utils/checks.py,sha256=mkDl_BTLZyjfhYbFVSG6xYmxhB2s7wsQ62ugnhspqOc,34707
+ultralytics/utils/checks.py,sha256=q64U5wKyejD-2W2fCPqJ0Oiaa4_4vq2pVxV9wp6lMz4,34707
 ultralytics/utils/dist.py,sha256=A9lDGtGefTjSVvVS38w86GOdbtLzNBDZuDGK0MT4PRI,4170
-ultralytics/utils/downloads.py,sha256=awaWFsx1k4wKESni5IgEmcAlAJVfKKpULhQmgmUhn2c,21916
+ultralytics/utils/downloads.py,sha256=A7r4LpWUojGkam9-VQ3Ylu-Cn1lAUGKyJE6VzwQbp7M,22016
 ultralytics/utils/errors.py,sha256=XT9Ru7ivoBgofK6PlnyigGoa7Fmf5nEhyHtnD-8TRXI,1584
 ultralytics/utils/export.py,sha256=LK-wlTlyb_zIKtSvOmfmvR70RcUU9Ct9UBDt5wn9_rY,9880
 ultralytics/utils/files.py,sha256=ZCbLGleiF0f-PqYfaxMFAWop88w7U1hpreHXl8b2ko0,8238
@@ -266,8 +266,8 @@ ultralytics/utils/callbacks/neptune.py,sha256=j8pecmlcsM8FGzLKWoBw5xUsi5t8E5HuxY
 ultralytics/utils/callbacks/raytune.py,sha256=S6Bq16oQDQ8BQgnZzA0zJHGN_BBr8iAM_WtGoLiEcwg,1283
 ultralytics/utils/callbacks/tensorboard.py,sha256=MDPBW7aDes-66OE6YqKXXvqA_EocjzEMHWGM-8z9vUQ,5281
 ultralytics/utils/callbacks/wb.py,sha256=Tm_-aRr2CN32MJkY9tylpMBJkb007-MSRNSQ7rDJ5QU,7521
-dgenerate_ultralytics_headless-8.3.178.dist-info/METADATA,sha256=qF1t4-NTjrCC-2ur8_Ay4sbps_4hp85OJEY4tgH_ZDM,38716
-dgenerate_ultralytics_headless-8.3.178.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-dgenerate_ultralytics_headless-8.3.178.dist-info/entry_points.txt,sha256=YM_wiKyTe9yRrsEfqvYolNO5ngwfoL4-NwgKzc8_7sI,93
-dgenerate_ultralytics_headless-8.3.178.dist-info/top_level.txt,sha256=XP49TwiMw4QGsvTLSYiJhz1xF_k7ev5mQ8jJXaXi45Q,12
-dgenerate_ultralytics_headless-8.3.178.dist-info/RECORD,,
+dgenerate_ultralytics_headless-8.3.180.dist-info/METADATA,sha256=-yv2HSf0JD7vFjkRFBIrwXoyHrcdbdQdhB5ocRW3_hk,38727
+dgenerate_ultralytics_headless-8.3.180.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+dgenerate_ultralytics_headless-8.3.180.dist-info/entry_points.txt,sha256=YM_wiKyTe9yRrsEfqvYolNO5ngwfoL4-NwgKzc8_7sI,93
+dgenerate_ultralytics_headless-8.3.180.dist-info/top_level.txt,sha256=XP49TwiMw4QGsvTLSYiJhz1xF_k7ev5mQ8jJXaXi45Q,12
+dgenerate_ultralytics_headless-8.3.180.dist-info/RECORD,,

ultralytics/__init__.py

@@ -1,6 +1,6 @@
 # Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license
 
-__version__ = "8.3.178"
+__version__ = "8.3.180"
 
 import os
 

ultralytics/cfg/__init__.py

@@ -954,7 +954,10 @@ def entrypoint(debug: str = "") -> None:
         from ultralytics import YOLO
 
         model = YOLO(model, task=task)
-
+        if "yoloe" in stem or "world" in stem:
+            cls_list = overrides.pop("classes", DEFAULT_CFG.classes)
+            if cls_list is not None and isinstance(cls_list, str):
+                model.set_classes(cls_list.split(","))  # convert "person, bus" -> ['person', ' bus'].
     # Task Update
     if task != model.task:
         if task:

ultralytics/cfg/datasets/VOC.yaml

@@ -87,7 +87,7 @@ download: |
       f"{url}VOCtest_06-Nov-2007.zip",  # 438MB, 4953 images
       f"{url}VOCtrainval_11-May-2012.zip",  # 1.95GB, 17126 images
   ]
-  download(urls, dir=dir / "images", curl=True, threads=3, exist_ok=True)  # download and unzip over existing (required)
+  download(urls, dir=dir / "images", threads=3, exist_ok=True)  # download and unzip over existing (required)
 
   # Convert
   path = dir / "images/VOCdevkit"

ultralytics/cfg/datasets/VisDrone.yaml

@@ -78,7 +78,7 @@ download: |
       "https://github.com/ultralytics/assets/releases/download/v0.0.0/VisDrone2019-DET-test-dev.zip",
       # "https://github.com/ultralytics/assets/releases/download/v0.0.0/VisDrone2019-DET-test-challenge.zip",
   ]
-  download(urls, dir=dir, curl=True, threads=4)
+  download(urls, dir=dir, threads=4)
 
   # Convert
   splits = {"VisDrone2019-DET-train": "train", "VisDrone2019-DET-val": "val", "VisDrone2019-DET-test-dev": "test"}

ultralytics/models/sam/predict.py

@@ -182,9 +182,8 @@ class Predictor(BasePredictor):
             **kwargs (Any): Additional keyword arguments.
 
         Returns:
-            pred_masks (np.ndarray): The output masks in shape (C, H, W), where C is the number of generated masks.
-            pred_scores (np.ndarray): An array of length C containing quality scores predicted by the model for each mask.
-            pred_logits (np.ndarray): Low-resolution logits of shape (C, H, W) for subsequent inference, where H=W=256.
+            pred_masks (torch.Tensor): The output masks in shape (C, H, W), where C is the number of generated masks.
+            pred_scores (torch.Tensor): An array of length C containing quality scores predicted by the model for each mask.
 
         Examples:
             >>> predictor = Predictor()
@@ -219,8 +218,8 @@ class Predictor(BasePredictor):
             multimask_output (bool): Flag to return multiple masks for ambiguous prompts.
 
         Returns:
-            pred_masks (np.ndarray): Output masks with shape (C, H, W), where C is the number of generated masks.
-            pred_scores (np.ndarray): Quality scores predicted by the model for each mask, with length C.
+            pred_masks (torch.Tensor): Output masks with shape (C, H, W), where C is the number of generated masks.
+            pred_scores (torch.Tensor): Quality scores predicted by the model for each mask, with length C.
 
         Examples:
             >>> predictor = Predictor()
@@ -230,7 +229,34 @@ class Predictor(BasePredictor):
         """
         features = self.get_im_features(im) if self.features is None else self.features
 
-        bboxes, points, labels, masks = self._prepare_prompts(im.shape[2:], bboxes, points, labels, masks)
+        prompts = self._prepare_prompts(im.shape[2:], self.batch[1][0].shape[:2], bboxes, points, labels, masks)
+        return self._inference_features(features, *prompts, multimask_output)
+
+    def _inference_features(
+        self,
+        features,
+        bboxes=None,
+        points=None,
+        labels=None,
+        masks=None,
+        multimask_output=False,
+    ):
+        """
+        Perform inference on image features using the SAM model.
+
+        Args:
+            features (torch.Tensor): Extracted image features with shape (B, C, H, W) from the SAM model image encoder.
+            bboxes (np.ndarray | List[List[float]] | None): Bounding boxes in XYXY format with shape (N, 4).
+            points (np.ndarray | List[List[float]] | None): Object location points with shape (N, 2), in pixels.
+            labels (np.ndarray | List[int] | None): Point prompt labels with shape (N,). 1 = foreground, 0 = background.
+            masks (List[np.ndarray] | np.ndarray | None): Masks for the objects, where each mask is a 2D array.
+            multimask_output (bool): Flag to return multiple masks for ambiguous prompts.
+            img_idx (int): Index of the image in the batch to process.
+
+        Returns:
+            pred_masks (torch.Tensor): Output masks with shape (C, H, W), where C is the number of generated masks.
+            pred_scores (torch.Tensor): Quality scores for each mask, with length C.
+        """
         points = (points, labels) if points is not None else None
         # Embed prompts
         sparse_embeddings, dense_embeddings = self.model.prompt_encoder(points=points, boxes=bboxes, masks=masks)
@@ -248,12 +274,13 @@ class Predictor(BasePredictor):
         # `d` could be 1 or 3 depends on `multimask_output`.
         return pred_masks.flatten(0, 1), pred_scores.flatten(0, 1)
 
-    def _prepare_prompts(self, dst_shape, bboxes=None, points=None, labels=None, masks=None):
+    def _prepare_prompts(self, dst_shape, src_shape, bboxes=None, points=None, labels=None, masks=None):
         """
         Prepare and transform the input prompts for processing based on the destination shape.
 
         Args:
-            dst_shape (tuple): The target shape (height, width) for the prompts.
+            dst_shape (Tuple[int, int]): The target shape (height, width) for the prompts.
+            src_shape (Tuple[int, int]): The source shape (height, width) of the input image.
             bboxes (np.ndarray | List | None): Bounding boxes in XYXY format with shape (N, 4).
             points (np.ndarray | List | None): Points indicating object locations with shape (N, 2) or (N, num_points, 2), in pixels.
             labels (np.ndarray | List | None): Point prompt labels with shape (N) or (N, num_points). 1 for foreground, 0 for background.
@@ -268,7 +295,6 @@ class Predictor(BasePredictor):
         Raises:
             AssertionError: If the number of points don't match the number of labels, in case labels were passed.
         """
-        src_shape = self.batch[1][0].shape[:2]
         r = 1.0 if self.segment_all else min(dst_shape[0] / src_shape[0], dst_shape[1] / src_shape[1])
         # Transform input prompts
         if points is not None:
@@ -543,7 +569,7 @@ class Predictor(BasePredictor):
             - The extracted features are stored in the `self.features` attribute for later use.
         """
         if self.model is None:
-            self.setup_model(model=None)
+            self.setup_model()
         self.setup_source(image)
         assert len(self.dataset) == 1, "`set_image` only supports setting one image!"
         for batch in self.dataset:
@@ -620,6 +646,53 @@ class Predictor(BasePredictor):
 
         return new_masks[keep].to(device=masks.device, dtype=masks.dtype), keep
 
+    @smart_inference_mode()
+    def inference_features(
+        self,
+        features,
+        src_shape,
+        dst_shape=None,
+        bboxes=None,
+        points=None,
+        labels=None,
+        masks=None,
+        multimask_output=False,
+    ):
+        """
+        Perform prompts preprocessing and inference on provided image features using the SAM model.
+
+        Args:
+            features (torch.Tensor | Dict[str, Any]): Extracted image features from the SAM/SAM2 model image encoder.
+            src_shape (Tuple[int, int]): The source shape (height, width) of the input image.
+            dst_shape (Tuple[int, int] | None): The target shape (height, width) for the prompts. If None, defaults to (imgsz, imgsz).
+            bboxes (np.ndarray | List[List[float]] | None): Bounding boxes in xyxy format with shape (N, 4).
+            points (np.ndarray | List[List[float]] | None): Points indicating object locations with shape (N, 2), in pixels.
+            labels (np.ndarray | List[int] | None): Point prompt labels with shape (N, ).
+            masks (List[np.ndarray] | np.ndarray | None): Masks for the objects, where each mask is a 2D array.
+            multimask_output (bool): Flag to return multiple masks for ambiguous prompts.
+
+        Returns:
+            pred_masks (torch.Tensor): The output masks in shape (C, H, W), where C is the number of generated masks.
+            pred_bboxes (torch.Tensor): Bounding boxes for each mask with shape (N, 6), where N is the number of boxes.
+                Each box is in xyxy format with additional columns for score and class.
+
+        Notes:
+            - The input features is a torch.Tensor of shape (B, C, H, W) if performing on SAM, or a Dict[str, Any] if performing on SAM2.
+        """
+        dst_shape = dst_shape or (self.args.imgsz, self.args.imgsz)
+        prompts = self._prepare_prompts(dst_shape, src_shape, bboxes, points, labels, masks)
+        pred_masks, pred_scores = self._inference_features(features, *prompts, multimask_output)
+        if len(pred_masks) == 0:
+            pred_masks, pred_bboxes = None, torch.zeros((0, 6), device=pred_masks.device)
+        else:
+            pred_masks = ops.scale_masks(pred_masks[None].float(), src_shape, padding=False)[0]
+            pred_masks = pred_masks > self.model.mask_threshold  # to bool
+            pred_bboxes = batched_mask_to_box(pred_masks)
+            # NOTE: SAM models do not return cls info. This `cls` here is just a placeholder for consistency.
+            cls = torch.arange(len(pred_masks), dtype=torch.int32, device=pred_masks.device)
+            pred_bboxes = torch.cat([pred_bboxes, pred_scores[:, None], cls[:, None]], dim=-1)
+        return pred_masks, pred_bboxes
+
 
 class SAM2Predictor(Predictor):
     """
@@ -663,80 +736,13 @@ class SAM2Predictor(Predictor):
 
         return build_sam(self.args.model)
 
-    def prompt_inference(
-        self,
-        im,
-        bboxes=None,
-        points=None,
-        labels=None,
-        masks=None,
-        multimask_output=False,
-        img_idx=-1,
-    ):
-        """
-        Perform image segmentation inference based on various prompts using SAM2 architecture.
-
-        This method leverages the Segment Anything Model 2 (SAM2) to generate segmentation masks for input images
-        based on provided prompts such as bounding boxes, points, or existing masks. It supports both single and
-        multi-object prediction scenarios.
-
-        Args:
-            im (torch.Tensor): Preprocessed input image tensor with shape (N, C, H, W).
-            bboxes (np.ndarray | List[List[float]] | None): Bounding boxes in XYXY format with shape (N, 4).
-            points (np.ndarray | List[List[float]] | None): Object location points with shape (N, 2), in pixels.
-            labels (np.ndarray | List[int] | None): Point prompt labels with shape (N,). 1 = foreground, 0 = background.
-            masks (np.ndarray | None): Low-resolution masks from previous predictions with shape (N, H, W).
-            multimask_output (bool): Flag to return multiple masks for ambiguous prompts.
-            img_idx (int): Index of the image in the batch to process.
-
-        Returns:
-            pred_masks (np.ndarray): Output masks with shape (C, H, W), where C is the number of generated masks.
-            pred_scores (np.ndarray): Quality scores for each mask, with length C.
-
-        Examples:
-            >>> predictor = SAM2Predictor(cfg)
-            >>> image = torch.rand(1, 3, 640, 640)
-            >>> bboxes = [[100, 100, 200, 200]]
-            >>> result = predictor(image, bboxes=bboxes)[0]
-            >>> print(f"Generated {result.masks.shape[0]} masks with average score {result.boxes.conf.mean():.2f}")
-
-        Notes:
-            - The method supports batched inference for multiple objects when points or bboxes are provided.
-            - Input prompts (bboxes, points) are automatically scaled to match the input image dimensions.
-            - When both bboxes and points are provided, they are merged into a single 'points' input for the model.
-        """
-        features = self.get_im_features(im) if self.features is None else self.features
-
-        points, labels, masks = self._prepare_prompts(im.shape[2:], bboxes, points, labels, masks)
-        points = (points, labels) if points is not None else None
-
-        sparse_embeddings, dense_embeddings = self.model.sam_prompt_encoder(
-            points=points,
-            boxes=None,
-            masks=masks,
-        )
-        # Predict masks
-        batched_mode = points is not None and points[0].shape[0] > 1  # multi object prediction
-        high_res_features = [feat_level[img_idx].unsqueeze(0) for feat_level in features["high_res_feats"]]
-        pred_masks, pred_scores, _, _ = self.model.sam_mask_decoder(
-            image_embeddings=features["image_embed"][img_idx].unsqueeze(0),
-            image_pe=self.model.sam_prompt_encoder.get_dense_pe(),
-            sparse_prompt_embeddings=sparse_embeddings,
-            dense_prompt_embeddings=dense_embeddings,
-            multimask_output=multimask_output,
-            repeat_image=batched_mode,
-            high_res_features=high_res_features,
-        )
-        # (N, d, H, W) --> (N*d, H, W), (N, d) --> (N*d, )
-        # `d` could be 1 or 3 depends on `multimask_output`.
-        return pred_masks.flatten(0, 1), pred_scores.flatten(0, 1)
-
-    def _prepare_prompts(self, dst_shape, bboxes=None, points=None, labels=None, masks=None):
+    def _prepare_prompts(self, dst_shape, src_shape, bboxes=None, points=None, labels=None, masks=None):
         """
         Prepare and transform the input prompts for processing based on the destination shape.
 
         Args:
-            dst_shape (tuple): The target shape (height, width) for the prompts.
+            dst_shape (Tuple[int, int]): The target shape (height, width) for the prompts.
+            src_shape (Tuple[int, int]): The source shape (height, width) of the input image.
             bboxes (np.ndarray | List | None): Bounding boxes in XYXY format with shape (N, 4).
             points (np.ndarray | List | None): Points indicating object locations with shape (N, 2) or (N, num_points, 2), in pixels.
             labels (np.ndarray | List | None): Point prompt labels with shape (N,) or (N, num_points). 1 for foreground, 0 for background.
@@ -750,7 +756,7 @@ class SAM2Predictor(Predictor):
         Raises:
             AssertionError: If the number of points don't match the number of labels, in case labels were passed.
         """
-        bboxes, points, labels, masks = super()._prepare_prompts(dst_shape, bboxes, points, labels, masks)
+        bboxes, points, labels, masks = super()._prepare_prompts(dst_shape, src_shape, bboxes, points, labels, masks)
         if bboxes is not None:
             bboxes = bboxes.view(-1, 2, 2)
             bbox_labels = torch.tensor([[2, 3]], dtype=torch.int32, device=bboxes.device).expand(len(bboxes), -1)
@@ -813,6 +819,54 @@ class SAM2Predictor(Predictor):
         ][::-1]
         return {"image_embed": feats[-1], "high_res_feats": feats[:-1]}
 
+    def _inference_features(
+        self,
+        features,
+        points=None,
+        labels=None,
+        masks=None,
+        multimask_output=False,
+        img_idx=-1,
+    ):
+        """
+        Perform inference on image features using the SAM2 model.
+
+        Args:
+            features (Dict[str, Any]): Extracted image information from the SAM2 model image encoder, it's a dictionary including:
+                image_embed (torch.Tensor): Image embedding with shape (B, C, H, W).
+                high_res_feats (List[torch.Tensor]): List of high-resolution feature maps from the backbone, each with shape (B, C, H, W).
+            points (np.ndarray | List[List[float]] | None): Object location points with shape (N, 2), in pixels.
+            labels (np.ndarray | List[int] | None): Point prompt labels with shape (N,). 1 = foreground, 0 = background.
+            masks (List[np.ndarray] | np.ndarray | None): Masks for the objects, where each mask is a 2D array.
+            multimask_output (bool): Flag to return multiple masks for ambiguous prompts.
+            img_idx (int): Index of the image in the batch to process.
+
+        Returns:
+            pred_masks (torch.Tensor): Output masks with shape (C, H, W), where C is the number of generated masks.
+            pred_scores (torch.Tensor): Quality scores for each mask, with length C.
+        """
+        points = (points, labels) if points is not None else None
+        sparse_embeddings, dense_embeddings = self.model.sam_prompt_encoder(
+            points=points,
+            boxes=None,
+            masks=masks,
+        )
+        # Predict masks
+        batched_mode = points is not None and points[0].shape[0] > 1  # multi object prediction
+        high_res_features = [feat_level[img_idx].unsqueeze(0) for feat_level in features["high_res_feats"]]
+        pred_masks, pred_scores, _, _ = self.model.sam_mask_decoder(
+            image_embeddings=features["image_embed"][img_idx].unsqueeze(0),
+            image_pe=self.model.sam_prompt_encoder.get_dense_pe(),
+            sparse_prompt_embeddings=sparse_embeddings,
+            dense_prompt_embeddings=dense_embeddings,
+            multimask_output=multimask_output,
+            repeat_image=batched_mode,
+            high_res_features=high_res_features,
+        )
+        # (N, d, H, W) --> (N*d, H, W), (N, d) --> (N*d, )
+        # `d` could be 1 or 3 depends on `multimask_output`.
+        return pred_masks.flatten(0, 1), pred_scores.flatten(0, 1)
+
 
 class SAM2VideoPredictor(SAM2Predictor):
     """
@@ -900,8 +954,8 @@ class SAM2VideoPredictor(SAM2Predictor):
             masks (np.ndarray, optional): Low-resolution masks from previous predictions shape (N,H,W). For SAM H=W=256.
 
         Returns:
-            pred_masks (np.ndarray): The output masks in shape CxHxW, where C is the number of generated masks.
-            pred_scores (np.ndarray): An array of length C containing quality scores predicted by the model for each mask.
+            pred_masks (torch.Tensor): The output masks in shape CxHxW, where C is the number of generated masks.
+            pred_scores (torch.Tensor): An array of length C containing quality scores predicted by the model for each mask.
         """
         # Override prompts if any stored in self.prompts
         bboxes = self.prompts.pop("bboxes", bboxes)
@@ -912,7 +966,9 @@ class SAM2VideoPredictor(SAM2Predictor):
         self.inference_state["im"] = im
         output_dict = self.inference_state["output_dict"]
         if len(output_dict["cond_frame_outputs"]) == 0:  # initialize prompts
-            points, labels, masks = self._prepare_prompts(im.shape[2:], bboxes, points, labels, masks)
+            points, labels, masks = self._prepare_prompts(
+                im.shape[2:], self.batch[1][0].shape[:2], bboxes, points, labels, masks
+            )
             if points is not None:
                 for i in range(len(points)):
                     self.add_new_prompts(obj_id=i, points=points[[i]], labels=labels[[i]], frame_idx=frame)
@@ -966,7 +1022,7 @@ class SAM2VideoPredictor(SAM2Predictor):
         the masks do not overlap, which can be useful for certain applications.
 
         Args:
-            preds (tuple): The predictions from the model.
+            preds (Tuple[torch.Tensor, torch.Tensor]): The predicted masks and scores from the model.
             img (torch.Tensor): The processed image tensor.
             orig_imgs (List[np.ndarray]): The original images before processing.
 

ultralytics/models/yolo/model.py

@@ -3,6 +3,8 @@
 from pathlib import Path
 from typing import Any, Dict, List, Optional, Union
 
+import torch
+
 from ultralytics.data.build import load_inference_source
 from ultralytics.engine.model import Model
 from ultralytics.models import yolo
@@ -315,7 +317,7 @@ class YOLOE(Model):
         assert isinstance(self.model, YOLOEModel)
         return self.model.get_vocab(names)
 
-    def set_classes(self, classes: List[str], embeddings) -> None:
+    def set_classes(self, classes: List[str], embeddings: Optional[torch.Tensor] = None) -> None:
         """
         Set the model's class names and embeddings for detection.
 
@@ -324,6 +326,8 @@ class YOLOE(Model):
             embeddings (torch.Tensor): Embeddings corresponding to the classes.
         """
         assert isinstance(self.model, YOLOEModel)
+        if embeddings is None:
+            embeddings = self.get_text_pe(classes)  # generate text embeddings if not provided
         self.model.set_classes(classes, embeddings)
         # Verify no background class is present
         assert " " not in classes

ultralytics/utils/checks.py

@@ -694,7 +694,7 @@ def collect_system_info():
         "GPU count": torch.cuda.device_count() if cuda else None,
         "CUDA": torch.version.cuda if cuda else None,
     }
-    LOGGER.info("\n" + "\n".join(f"{k:<20}{v}" for k, v in info_dict.items()) + "\n")
+    LOGGER.info("\n" + "\n".join(f"{k:<23}{v}" for k, v in info_dict.items()) + "\n")
 
     package_info = {}
     for r in parse_requirements(package="ultralytics"):
@@ -705,7 +705,7 @@ def collect_system_info():
             current = "(not installed)"
             is_met = "❌ "
         package_info[r.name] = f"{is_met}{current}{r.specifier}"
-        LOGGER.info(f"{r.name:<20}{package_info[r.name]}")
+        LOGGER.info(f"{r.name:<23}{package_info[r.name]}")
 
     info_dict["Package Info"] = package_info
 

ultralytics/utils/downloads.py

@@ -501,7 +501,9 @@ def download(
     """
     dir = Path(dir)
     dir.mkdir(parents=True, exist_ok=True)  # make directory
+    urls = [url] if isinstance(url, (str, Path)) else url
     if threads > 1:
+        LOGGER.info(f"Downloading {len(urls)} file(s) with {threads} threads to {dir}...")
         with ThreadPool(threads) as pool:
             pool.map(
                 lambda x: safe_download(
@@ -512,12 +514,12 @@ def download(
                     curl=curl,
                     retry=retry,
                     exist_ok=exist_ok,
-                    progress=threads <= 1,
+                    progress=True,
                 ),
-                zip(url, repeat(dir)),
+                zip(urls, repeat(dir)),
             )
             pool.close()
             pool.join()
     else:
-        for u in [url] if isinstance(url, (str, Path)) else url:
+        for u in urls:
             safe_download(url=u, dir=dir, unzip=unzip, delete=delete, curl=curl, retry=retry, exist_ok=exist_ok)