ultralytics 8.3.151__tar.gz → 8.3.152__tar.gz

{ultralytics-8.3.151/ultralytics.egg-info → ultralytics-8.3.152}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: ultralytics
-Version: 8.3.151
+Version: 8.3.152
 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-8.3.151 → ultralytics-8.3.152}/ultralytics/__init__.py

@@ -1,6 +1,6 @@
 # Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license
 
-__version__ = "8.3.151"
+__version__ = "8.3.152"
 
 import os
 

{ultralytics-8.3.151 → ultralytics-8.3.152}/ultralytics/engine/model.py

@@ -293,7 +293,7 @@ class Model(torch.nn.Module):
 
         if str(weights).rpartition(".")[-1] == "pt":
             self.model, self.ckpt = attempt_load_one_weight(weights)
-            self.task = self.model.args["task"]
+            self.task = self.model.task
             self.overrides = self.model.args = self._reset_ckpt_args(self.model.args)
             self.ckpt_path = self.model.pt_path
         else:

{ultralytics-8.3.151 → ultralytics-8.3.152}/ultralytics/engine/trainer.py

@@ -456,6 +456,7 @@ class BaseTrainer:
 
                 # Validation
                 if self.args.val or final_epoch or self.stopper.possible_stop or self.stop:
+                    self._clear_memory(threshold=0.5)  # prevent VRAM spike
                     self.metrics, self.fitness = self.validate()
                 self.save_metrics(metrics={**self.label_loss_items(self.tloss), **self.metrics, **self.lr})
                 self.stop |= self.stopper(epoch + 1, self.fitness) or final_epoch
@@ -478,8 +479,7 @@ class BaseTrainer:
                 self.scheduler.last_epoch = self.epoch  # do not move
                 self.stop |= epoch >= self.epochs  # stop if exceeded epochs
             self.run_callbacks("on_fit_epoch_end")
-            if self._get_memory(fraction=True) > 0.5:
-                self._clear_memory()  # clear if memory utilization > 50%
+            self._clear_memory(0.5)  # clear if memory utilization > 50%
 
             # Early Stopping
             if RANK != -1:  # if DDP training
@@ -525,8 +525,12 @@ class BaseTrainer:
                 total = torch.cuda.get_device_properties(self.device).total_memory
         return ((memory / total) if total > 0 else 0) if fraction else (memory / 2**30)
 
-    def _clear_memory(self):
+    def _clear_memory(self, threshold: float = None):
         """Clear accelerator memory by calling garbage collector and emptying cache."""
+        if threshold:
+            assert 0 <= threshold <= 1, "Threshold must be between 0 and 1."
+            if self._get_memory(fraction=True) <= threshold:
+                return
         gc.collect()
         if self.device.type == "mps":
             torch.mps.empty_cache()

{ultralytics-8.3.151 → ultralytics-8.3.152}/ultralytics/nn/tasks.py

@@ -1505,7 +1505,7 @@ def attempt_load_weights(weights, device=None, inplace=True, fuse=False):
         # Model compatibility updates
         model.args = args  # attach args to model
         model.pt_path = w  # attach *.pt file path to model
-        model.task = guess_model_task(model)
+        model.task = getattr(model, "task", guess_model_task(model))
         if not hasattr(model, "stride"):
             model.stride = torch.tensor([32.0])
 
@@ -1553,7 +1553,7 @@ def attempt_load_one_weight(weight, device=None, inplace=True, fuse=False):
     # Model compatibility updates
     model.args = {k: v for k, v in args.items() if k in DEFAULT_CFG_KEYS}  # attach args to model
     model.pt_path = weight  # attach *.pt file path to model
-    model.task = guess_model_task(model)
+    model.task = getattr(model, "task", guess_model_task(model))
     if not hasattr(model, "stride"):
         model.stride = torch.tensor([32.0])
 

{ultralytics-8.3.151 → ultralytics-8.3.152}/ultralytics/utils/metrics.py

@@ -365,19 +365,19 @@ class ConfusionMatrix(DataExportMixin):
         if gt_cls.shape[0] == 0:  # Check if labels is empty
             if detections is not None:
                 detections = detections[detections[:, 4] > self.conf]
-                detection_classes = detections[:, 5].int()
+                detection_classes = detections[:, 5].int().tolist()
                 for dc in detection_classes:
                     self.matrix[dc, self.nc] += 1  # false positives
             return
         if detections is None:
-            gt_classes = gt_cls.int()
+            gt_classes = gt_cls.int().tolist()
             for gc in gt_classes:
                 self.matrix[self.nc, gc] += 1  # background FN
             return
 
         detections = detections[detections[:, 4] > self.conf]
-        gt_classes = gt_cls.int()
-        detection_classes = detections[:, 5].int()
+        gt_classes = gt_cls.int().tolist()
+        detection_classes = detections[:, 5].int().tolist()
         is_obb = detections.shape[1] == 7 and gt_bboxes.shape[1] == 5  # with additional `angle` dimension
         iou = (
             batch_probiou(gt_bboxes, torch.cat([detections[:, :4], detections[:, -1:]], dim=-1))
@@ -401,7 +401,7 @@ class ConfusionMatrix(DataExportMixin):
         for i, gc in enumerate(gt_classes):
             j = m0 == i
             if n and sum(j) == 1:
-                self.matrix[detection_classes[m1[j]], gc] += 1  # correct
+                self.matrix[detection_classes[m1[j].item()], gc] += 1  # correct
             else:
                 self.matrix[self.nc, gc] += 1  # true background
 
@@ -463,6 +463,7 @@ class ConfusionMatrix(DataExportMixin):
             im = ax.imshow(array, cmap="Blues", vmin=0.0, interpolation="none")
             ax.xaxis.set_label_position("bottom")
             if nc < 30:  # Add score for each cell of confusion matrix
+                color_threshold = 0.45 * (1 if normalize else np.nanmax(array))  # text color threshold
                 for i, row in enumerate(array[:nc]):
                     for j, val in enumerate(row[:nc]):
                         val = array[i, j]
@@ -475,7 +476,7 @@ class ConfusionMatrix(DataExportMixin):
                             ha="center",
                             va="center",
                             fontsize=10,
-                            color="white" if val > (0.7 if normalize else 2) else "black",
+                            color="white" if val > color_threshold else "black",
                         )
             cbar = fig.colorbar(im, ax=ax, fraction=0.046, pad=0.05)
         title = "Confusion Matrix" + " Normalized" * normalize

{ultralytics-8.3.151 → ultralytics-8.3.152}/ultralytics/utils/ops.py

@@ -404,8 +404,12 @@ def scale_image(masks, im0_shape, ratio_pad=None):
         pad = (im1_shape[1] - im0_shape[1] * gain) / 2, (im1_shape[0] - im0_shape[0] * gain) / 2  # wh padding
     else:
         pad = ratio_pad[1]
-    top, left = int(pad[1]), int(pad[0])  # y, x
-    bottom, right = int(im1_shape[0] - pad[1]), int(im1_shape[1] - pad[0])
+
+    top, left = (int(round(pad[1] - 0.1)), int(round(pad[0] - 0.1)))
+    bottom, right = (
+        im1_shape[0] - int(round(pad[1] + 0.1)),
+        im1_shape[1] - int(round(pad[0] + 0.1)),
+    )
 
     if len(masks.shape) < 2:
         raise ValueError(f'"len of masks shape" should be 2 or 3, but got {len(masks.shape)}')
@@ -756,8 +760,11 @@ def scale_masks(masks, shape, padding: bool = True):
     if padding:
         pad[0] /= 2
         pad[1] /= 2
-    top, left = (int(pad[1]), int(pad[0])) if padding else (0, 0)  # y, x
-    bottom, right = (int(mh - pad[1]), int(mw - pad[0]))
+    top, left = (int(round(pad[1] - 0.1)), int(round(pad[0] - 0.1))) if padding else (0, 0)  # y, x
+    bottom, right = (
+        mh - int(round(pad[1] + 0.1)),
+        mw - int(round(pad[0] + 0.1)),
+    )
     masks = masks[..., top:bottom, left:right]
 
     masks = F.interpolate(masks, shape, mode="bilinear", align_corners=False)  # NCHW

{ultralytics-8.3.151 → ultralytics-8.3.152/ultralytics.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: ultralytics
-Version: 8.3.151
+Version: 8.3.152
 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>