s,np.bool\b,bool,

Summary:
X-link: facebookresearch/Detectron#1047

X-link: facebook/openzl#582

`np.bool` and `bool` are identical, but the former is removed from Numpy-1.24.X.

Differential Revision: D49196618

Author: igorsugak

mmdet/core/evaluation/mean_ap.py

@@ -84,8 +84,8 @@ def tpfp_imagenet(
     # an indicator of ignored gts
     gt_ignore_inds = np.concatenate(
         (
-            np.zeros(gt_bboxes.shape[0], dtype=np.bool),
-            np.ones(gt_bboxes_ignore.shape[0], dtype=np.bool),
+            np.zeros(gt_bboxes.shape[0], dtype=bool),
+            np.ones(gt_bboxes_ignore.shape[0], dtype=bool),
         )
     )
     # stack gt_bboxes and gt_bboxes_ignore for convenience
@@ -179,8 +179,8 @@ def tpfp_default(
     # an indicator of ignored gts
     gt_ignore_inds = np.concatenate(
         (
-            np.zeros(gt_bboxes.shape[0], dtype=np.bool),
-            np.ones(gt_bboxes_ignore.shape[0], dtype=np.bool),
+            np.zeros(gt_bboxes.shape[0], dtype=bool),
+            np.ones(gt_bboxes_ignore.shape[0], dtype=bool),
         )
     )
     # stack gt_bboxes and gt_bboxes_ignore for convenience

mmdet/core/mask/structures.py

@@ -860,5 +860,5 @@ def polygon_to_bitmap(polygons, height, width):
     """
     rles = maskUtils.frPyObjects(polygons, height, width)
     rle = maskUtils.merge(rles)
-    bitmap_mask = maskUtils.decode(rle).astype(np.bool)
+    bitmap_mask = maskUtils.decode(rle).astype(bool)
     return bitmap_mask

mmdet/models/detectors/pa_predictor.py

@@ -51,7 +51,7 @@ def palette2filter(palette, neighbor_sizes=None, bidirection=True):
     """
 
     def generate_cross_filter(number_of_neighbors, dist):
-        cross_filter = np.zeros((number_of_neighbors,) + palette.shape, dtype=np.bool)
+        cross_filter = np.zeros((number_of_neighbors,) + palette.shape, dtype=bool)
         cross_filter[0, dist:, :] = np.logical_or(palette[dist:, :], palette[:-dist, :])
         cross_filter[1, :, dist:] = np.logical_or(palette[:, dist:], palette[:, :-dist])
         cross_filter[2, dist:, dist:] = np.logical_or(
@@ -81,7 +81,7 @@ def generate_cross_filter(number_of_neighbors, dist):
     number_of_span = len(neighbor_sizes)
     number_of_neighbors_per_span = 8 if bidirection else 4
     potential_filter = np.zeros(
-        (number_of_neighbors_per_span * number_of_span,) + palette.shape, dtype=np.bool
+        (number_of_neighbors_per_span * number_of_span,) + palette.shape, dtype=bool
     )
     for neighbor_idx, dist in enumerate(neighbor_sizes):
         offset = neighbor_idx * number_of_neighbors_per_span
@@ -103,7 +103,7 @@ def palette2weight(
     """
 
     def generate_cross_weight(number_of_neighbors, dist):
-        cross_filter = np.zeros((number_of_neighbors,) + palette.shape, dtype=np.bool)
+        cross_filter = np.zeros((number_of_neighbors,) + palette.shape, dtype=bool)
         cross_filter[0, dist:, :] = np.maximum(palette[dist:, :], palette[:-dist, :])
         cross_filter[1, :, dist:] = np.maximum(palette[:, dist:], palette[:, :-dist])
         cross_filter[2, dist:, dist:] = np.maximum(
@@ -152,7 +152,7 @@ def generate_cross_weight(number_of_neighbors, dist):
     number_of_span = len(neighbor_sizes)
     number_of_neighbors_per_span = 8 if bidirection else 4
     potential_weight = np.zeros(
-        (number_of_neighbors_per_span * number_of_span,) + palette.shape, dtype=np.bool
+        (number_of_neighbors_per_span * number_of_span,) + palette.shape, dtype=bool
     )
     for neighbor_idx, dist in enumerate(neighbor_sizes):
         offset = neighbor_idx * number_of_neighbors_per_span

pa_lib/evaluate_helper.py

@@ -61,8 +61,8 @@ def rank_by_variance(image, masks, weight_by_size=False, nms=1.0):
 def compute_iou(annotation, segmentation, mask_threshold=0.0):
     if type(annotation) == torch.Tensor:
         annotation = annotation.numpy()
-    annotation = annotation.astype(np.bool)
-    segmentation = (segmentation > mask_threshold).astype(np.bool)
+    annotation = annotation.astype(bool)
+    segmentation = (segmentation > mask_threshold).astype(bool)
 
     if np.isclose(np.sum(annotation), 0) and np.isclose(np.sum(segmentation), 0):
         return 1