ultralytics 8.3.88__py3-none-any.whl → 8.3.90__py3-none-any.whl

ultralytics/utils/tal.py

@@ -21,6 +21,7 @@ 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.
@@ -39,23 +40,25 @@ class TaskAlignedAssigner(nn.Module):
     @torch.no_grad()
     def forward(self, pd_scores, pd_bboxes, anc_points, gt_labels, gt_bboxes, mask_gt):
         """
-        Compute the task-aligned assignment. Reference code is available at
-        https://github.com/Nioolek/PPYOLOE_pytorch/blob/master/ppyoloe/assigner/tal_assigner.py.
+        Compute the task-aligned assignment.
 
         Args:
-            pd_scores (Tensor): shape(bs, num_total_anchors, num_classes)
-            pd_bboxes (Tensor): shape(bs, num_total_anchors, 4)
-            anc_points (Tensor): shape(num_total_anchors, 2)
-            gt_labels (Tensor): shape(bs, n_max_boxes, 1)
-            gt_bboxes (Tensor): shape(bs, n_max_boxes, 4)
-            mask_gt (Tensor): shape(bs, n_max_boxes, 1)
+            pd_scores (torch.Tensor): Predicted classification scores with shape (bs, num_total_anchors, num_classes).
+            pd_bboxes (torch.Tensor): Predicted bounding boxes with shape (bs, num_total_anchors, 4).
+            anc_points (torch.Tensor): Anchor points with shape (num_total_anchors, 2).
+            gt_labels (torch.Tensor): Ground truth labels with shape (bs, n_max_boxes, 1).
+            gt_bboxes (torch.Tensor): Ground truth boxes with shape (bs, n_max_boxes, 4).
+            mask_gt (torch.Tensor): Mask for valid ground truth boxes with shape (bs, n_max_boxes, 1).
 
         Returns:
-            target_labels (Tensor): shape(bs, num_total_anchors)
-            target_bboxes (Tensor): shape(bs, num_total_anchors, 4)
-            target_scores (Tensor): shape(bs, num_total_anchors, num_classes)
-            fg_mask (Tensor): shape(bs, num_total_anchors)
-            target_gt_idx (Tensor): shape(bs, num_total_anchors)
+            target_labels (torch.Tensor): Target labels with shape (bs, num_total_anchors).
+            target_bboxes (torch.Tensor): Target bounding boxes with shape (bs, num_total_anchors, 4).
+            target_scores (torch.Tensor): Target scores with shape (bs, num_total_anchors, num_classes).
+            fg_mask (torch.Tensor): Foreground mask with shape (bs, num_total_anchors).
+            target_gt_idx (torch.Tensor): Target ground truth indices with shape (bs, num_total_anchors).
+
+        References:
+            https://github.com/Nioolek/PPYOLOE_pytorch/blob/master/ppyoloe/assigner/tal_assigner.py
         """
         self.bs = pd_scores.shape[0]
         self.n_max_boxes = gt_bboxes.shape[1]
@@ -81,23 +84,22 @@ class TaskAlignedAssigner(nn.Module):
 
     def _forward(self, pd_scores, pd_bboxes, anc_points, gt_labels, gt_bboxes, mask_gt):
         """
-        Compute the task-aligned assignment. Reference code is available at
-        https://github.com/Nioolek/PPYOLOE_pytorch/blob/master/ppyoloe/assigner/tal_assigner.py.
+        Compute the task-aligned assignment.
 
         Args:
-            pd_scores (Tensor): shape(bs, num_total_anchors, num_classes)
-            pd_bboxes (Tensor): shape(bs, num_total_anchors, 4)
-            anc_points (Tensor): shape(num_total_anchors, 2)
-            gt_labels (Tensor): shape(bs, n_max_boxes, 1)
-            gt_bboxes (Tensor): shape(bs, n_max_boxes, 4)
-            mask_gt (Tensor): shape(bs, n_max_boxes, 1)
+            pd_scores (torch.Tensor): Predicted classification scores with shape (bs, num_total_anchors, num_classes).
+            pd_bboxes (torch.Tensor): Predicted bounding boxes with shape (bs, num_total_anchors, 4).
+            anc_points (torch.Tensor): Anchor points with shape (num_total_anchors, 2).
+            gt_labels (torch.Tensor): Ground truth labels with shape (bs, n_max_boxes, 1).
+            gt_bboxes (torch.Tensor): Ground truth boxes with shape (bs, n_max_boxes, 4).
+            mask_gt (torch.Tensor): Mask for valid ground truth boxes with shape (bs, n_max_boxes, 1).
 
         Returns:
-            target_labels (Tensor): shape(bs, num_total_anchors)
-            target_bboxes (Tensor): shape(bs, num_total_anchors, 4)
-            target_scores (Tensor): shape(bs, num_total_anchors, num_classes)
-            fg_mask (Tensor): shape(bs, num_total_anchors)
-            target_gt_idx (Tensor): shape(bs, num_total_anchors)
+            target_labels (torch.Tensor): Target labels with shape (bs, num_total_anchors).
+            target_bboxes (torch.Tensor): Target bounding boxes with shape (bs, num_total_anchors, 4).
+            target_scores (torch.Tensor): Target scores with shape (bs, num_total_anchors, num_classes).
+            fg_mask (torch.Tensor): Foreground mask with shape (bs, num_total_anchors).
+            target_gt_idx (torch.Tensor): Target ground truth indices with shape (bs, num_total_anchors).
         """
         mask_pos, align_metric, overlaps = self.get_pos_mask(
             pd_scores, pd_bboxes, gt_labels, gt_bboxes, anc_points, mask_gt
@@ -118,7 +120,22 @@ class TaskAlignedAssigner(nn.Module):
         return target_labels, target_bboxes, target_scores, fg_mask.bool(), target_gt_idx
 
     def get_pos_mask(self, pd_scores, pd_bboxes, gt_labels, gt_bboxes, anc_points, mask_gt):
-        """Get in_gts mask, (b, max_num_obj, h*w)."""
+        """
+        Get positive mask for each ground truth box.
+
+        Args:
+            pd_scores (torch.Tensor): Predicted classification scores with shape (bs, num_total_anchors, num_classes).
+            pd_bboxes (torch.Tensor): Predicted bounding boxes with shape (bs, num_total_anchors, 4).
+            gt_labels (torch.Tensor): Ground truth labels with shape (bs, n_max_boxes, 1).
+            gt_bboxes (torch.Tensor): Ground truth boxes with shape (bs, n_max_boxes, 4).
+            anc_points (torch.Tensor): Anchor points with shape (num_total_anchors, 2).
+            mask_gt (torch.Tensor): Mask for valid ground truth boxes with shape (bs, n_max_boxes, 1).
+
+        Returns:
+            mask_pos (torch.Tensor): Positive mask with shape (bs, max_num_obj, h*w).
+            align_metric (torch.Tensor): Alignment metric with shape (bs, max_num_obj, h*w).
+            overlaps (torch.Tensor): Overlaps between predicted and ground truth boxes with shape (bs, max_num_obj, h*w).
+        """
         mask_in_gts = self.select_candidates_in_gts(anc_points, gt_bboxes)
         # Get anchor_align metric, (b, max_num_obj, h*w)
         align_metric, overlaps = self.get_box_metrics(pd_scores, pd_bboxes, gt_labels, gt_bboxes, mask_in_gts * mask_gt)
@@ -130,7 +147,20 @@ class TaskAlignedAssigner(nn.Module):
         return mask_pos, align_metric, overlaps
 
     def get_box_metrics(self, pd_scores, pd_bboxes, gt_labels, gt_bboxes, mask_gt):
-        """Compute alignment metric given predicted and ground truth bounding boxes."""
+        """
+        Compute alignment metric given predicted and ground truth bounding boxes.
+
+        Args:
+            pd_scores (torch.Tensor): Predicted classification scores with shape (bs, num_total_anchors, num_classes).
+            pd_bboxes (torch.Tensor): Predicted bounding boxes with shape (bs, num_total_anchors, 4).
+            gt_labels (torch.Tensor): Ground truth labels with shape (bs, n_max_boxes, 1).
+            gt_bboxes (torch.Tensor): Ground truth boxes with shape (bs, n_max_boxes, 4).
+            mask_gt (torch.Tensor): Mask for valid ground truth boxes with shape (bs, n_max_boxes, h*w).
+
+        Returns:
+            align_metric (torch.Tensor): Alignment metric combining classification and localization.
+            overlaps (torch.Tensor): IoU overlaps between predicted and ground truth boxes.
+        """
         na = pd_bboxes.shape[-2]
         mask_gt = mask_gt.bool()  # b, max_num_obj, h*w
         overlaps = torch.zeros([self.bs, self.n_max_boxes, na], dtype=pd_bboxes.dtype, device=pd_bboxes.device)
@@ -151,7 +181,16 @@ class TaskAlignedAssigner(nn.Module):
         return align_metric, overlaps
 
     def iou_calculation(self, gt_bboxes, pd_bboxes):
-        """IoU calculation for horizontal bounding boxes."""
+        """
+        Calculate IoU for horizontal bounding boxes.
+
+        Args:
+            gt_bboxes (torch.Tensor): Ground truth boxes.
+            pd_bboxes (torch.Tensor): Predicted boxes.
+
+        Returns:
+            (torch.Tensor): IoU values between each pair of boxes.
+        """
         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):
@@ -159,16 +198,16 @@ class TaskAlignedAssigner(nn.Module):
         Select the top-k candidates based on the given metrics.
 
         Args:
-            metrics (Tensor): A tensor of shape (b, max_num_obj, h*w), where b is the batch size,
+            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 (Tensor): An optional boolean tensor of shape (b, max_num_obj, topk), where
+            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.
 
         Returns:
-            (Tensor): A tensor of shape (b, max_num_obj, h*w) containing the selected top-k candidates.
+            (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)
@@ -183,7 +222,6 @@ class TaskAlignedAssigner(nn.Module):
         for k in range(self.topk):
             # Expand topk_idxs for each value of k and add 1 at the specified positions
             count_tensor.scatter_add_(-1, topk_idxs[:, :, k : k + 1], ones)
-        # count_tensor.scatter_add_(-1, topk_idxs, torch.ones_like(topk_idxs, dtype=torch.int8, device=topk_idxs.device))
         # Filter invalid bboxes
         count_tensor.masked_fill_(count_tensor > 1, 0)
 
@@ -194,24 +232,21 @@ class TaskAlignedAssigner(nn.Module):
         Compute target labels, target bounding boxes, and target scores for the positive anchor points.
 
         Args:
-            gt_labels (Tensor): Ground truth labels of shape (b, max_num_obj, 1), where b is the
+            gt_labels (torch.Tensor): Ground truth labels of shape (b, max_num_obj, 1), where b is the
                                 batch size and max_num_obj is the maximum number of objects.
-            gt_bboxes (Tensor): Ground truth bounding boxes of shape (b, max_num_obj, 4).
-            target_gt_idx (Tensor): Indices of the assigned ground truth objects for positive
+            gt_bboxes (torch.Tensor): Ground truth bounding boxes of shape (b, max_num_obj, 4).
+            target_gt_idx (torch.Tensor): Indices of the assigned ground truth objects for positive
                                     anchor points, with shape (b, h*w), where h*w is the total
                                     number of anchor points.
-            fg_mask (Tensor): A boolean tensor of shape (b, h*w) indicating the positive
+            fg_mask (torch.Tensor): A boolean tensor of shape (b, h*w) indicating the positive
                               (foreground) anchor points.
 
         Returns:
-            (Tuple[Tensor, Tensor, Tensor]): A tuple containing the following tensors:
-                - target_labels (Tensor): Shape (b, h*w), containing the target labels for
-                                          positive anchor points.
-                - target_bboxes (Tensor): Shape (b, h*w, 4), containing the target bounding boxes
-                                          for positive anchor points.
-                - target_scores (Tensor): Shape (b, h*w, num_classes), containing the target scores
-                                          for positive anchor points, where num_classes is the number
-                                          of object classes.
+            target_labels (torch.Tensor): Shape (b, h*w), containing the target labels for positive anchor points.
+            target_bboxes (torch.Tensor): Shape (b, h*w, 4), containing the target bounding boxes for positive
+                                          anchor points.
+            target_scores (torch.Tensor): Shape (b, h*w, num_classes), containing the target scores for positive
+                                          anchor points.
         """
         # Assigned target labels, (b, 1)
         batch_ind = torch.arange(end=self.bs, dtype=torch.int64, device=gt_labels.device)[..., None]
@@ -258,7 +293,6 @@ class TaskAlignedAssigner(nn.Module):
         bs, n_boxes, _ = gt_bboxes.shape
         lt, rb = gt_bboxes.view(-1, 1, 4).chunk(2, 2)  # left-top, right-bottom
         bbox_deltas = torch.cat((xy_centers[None] - lt, rb - xy_centers[None]), dim=2).view(bs, n_boxes, n_anchors, -1)
-        # return (bbox_deltas.min(3)[0] > eps).to(gt_bboxes.dtype)
         return bbox_deltas.amin(3).gt_(eps)
 
     @staticmethod
@@ -275,9 +309,6 @@ class TaskAlignedAssigner(nn.Module):
             target_gt_idx (torch.Tensor): Indices of assigned ground truths, shape (b, h*w).
             fg_mask (torch.Tensor): Foreground mask, shape (b, h*w).
             mask_pos (torch.Tensor): Updated positive mask, shape (b, n_max_boxes, h*w).
-
-        Note:
-            b: batch size, h: height, w: width.
         """
         # Convert (b, n_max_boxes, h*w) -> (b, h*w)
         fg_mask = mask_pos.sum(-2)
@@ -299,7 +330,7 @@ class RotatedTaskAlignedAssigner(TaskAlignedAssigner):
     """Assigns ground-truth objects to rotated bounding boxes using a task-aligned metric."""
 
     def iou_calculation(self, gt_bboxes, pd_bboxes):
-        """IoU calculation for rotated bounding boxes."""
+        """Calculate IoU for rotated bounding boxes."""
         return probiou(gt_bboxes, pd_bboxes).squeeze(-1).clamp_(0)
 
     @staticmethod
@@ -308,11 +339,11 @@ class RotatedTaskAlignedAssigner(TaskAlignedAssigner):
         Select the positive anchor center in gt for rotated bounding boxes.
 
         Args:
-            xy_centers (Tensor): shape(h*w, 2)
-            gt_bboxes (Tensor): shape(b, n_boxes, 5)
+            xy_centers (torch.Tensor): Anchor center coordinates with shape (h*w, 2).
+            gt_bboxes (torch.Tensor): Ground truth bounding boxes with shape (b, n_boxes, 5).
 
         Returns:
-            (Tensor): shape(b, n_boxes, h*w)
+            (torch.Tensor): Boolean mask of positive anchors with shape (b, n_boxes, h*w).
         """
         # (b, n_boxes, 5) --> (b, n_boxes, 4, 2)
         corners = xywhr2xyxyxyxy(gt_bboxes)
@@ -368,13 +399,13 @@ def dist2rbox(pred_dist, pred_angle, anchor_points, dim=-1):
     Decode predicted rotated bounding box coordinates from anchor points and distribution.
 
     Args:
-        pred_dist (torch.Tensor): Predicted rotated distance, shape (bs, h*w, 4).
-        pred_angle (torch.Tensor): Predicted angle, shape (bs, h*w, 1).
-        anchor_points (torch.Tensor): Anchor points, shape (h*w, 2).
+        pred_dist (torch.Tensor): Predicted rotated distance with shape (bs, h*w, 4).
+        pred_angle (torch.Tensor): Predicted angle with shape (bs, h*w, 1).
+        anchor_points (torch.Tensor): Anchor points with shape (h*w, 2).
         dim (int, optional): Dimension along which to split. Defaults to -1.
 
     Returns:
-        (torch.Tensor): Predicted rotated bounding boxes, shape (bs, h*w, 4).
+        (torch.Tensor): Predicted rotated bounding boxes with shape (bs, h*w, 4).
     """
     lt, rb = pred_dist.split(2, dim=dim)
     cos, sin = torch.cos(pred_angle), torch.sin(pred_angle)