fix

mmdet/core/bbox/samplers/base_sampler.py

@@ -58,9 +58,8 @@ class BaseSampler(metaclass=ABCMeta):
             gt_flags = torch.cat([gt_ones, gt_flags])
 
         num_expected_pos = int(self.num * self.pos_fraction)
-        kwargs.update(dict(bboxes=bboxes))
-        pos_inds = self.pos_sampler._sample_pos(assign_result,
-                                                num_expected_pos, **kwargs)
+        pos_inds = self.pos_sampler._sample_pos(
+            assign_result, num_expected_pos, bboxes=bboxes, **kwargs)
         # We found that sampled indices have duplicated items occasionally.
         # (may be a bug of PyTorch)
         pos_inds = pos_inds.unique()
@@ -71,8 +70,8 @@ class BaseSampler(metaclass=ABCMeta):
             neg_upper_bound = int(self.neg_pos_ub * _pos)
             if num_expected_neg > neg_upper_bound:
                 num_expected_neg = neg_upper_bound
-        neg_inds = self.neg_sampler._sample_neg(assign_result,
-                                                num_expected_neg, **kwargs)
+        neg_inds = self.neg_sampler._sample_neg(
+            assign_result, num_expected_neg, bboxes=bboxes, **kwargs)
         neg_inds = neg_inds.unique()
 
         return SamplingResult(pos_inds, neg_inds, bboxes, gt_bboxes,

mmdet/core/bbox/samplers/ohem_sampler.py

@@ -9,17 +9,16 @@ class OHEMSampler(BaseSampler):
     def __init__(self,
                  num,
                  pos_fraction,
-                 neg_pos_ub,
-                 add_gt_as_proposals,
                  context,
+                 neg_pos_ub=-1,
+                 add_gt_as_proposals=True,
                  **kwargs):
         super(OHEMSampler, self).__init__(num, pos_fraction, neg_pos_ub,
-                                          add_gt_as_proposals, **kwargs)
+                                          add_gt_as_proposals)
         self.bbox_roi_extractor = context.bbox_roi_extractor
         self.bbox_head = context.bbox_head
 
     def hard_mining(self, inds, num_expected, bboxes, labels, feats):
-        # hard mining from the gallery.
         with torch.no_grad():
             rois = bbox2roi([bboxes])
             bbox_feats = self.bbox_roi_extractor(
@@ -42,7 +41,7 @@ class OHEMSampler(BaseSampler):
                     bboxes=None,
                     feats=None,
                     **kwargs):
-        # Hard sample some positive samples
+        # Sample some hard positive samples
         pos_inds = torch.nonzero(assign_result.gt_inds > 0)
         if pos_inds.numel() != 0:
             pos_inds = pos_inds.squeeze(1)
@@ -58,7 +57,7 @@ class OHEMSampler(BaseSampler):
                     bboxes=None,
                     feats=None,
                     **kwargs):
-        # Hard sample some negative samples
+        # Sample some hard negative samples
         neg_inds = torch.nonzero(assign_result.gt_inds == 0)
         if neg_inds.numel() != 0:
             neg_inds = neg_inds.squeeze(1)

mmdet/core/bbox/samplers/random_sampler.py

@@ -13,11 +13,7 @@ class RandomSampler(BaseSampler):
                  add_gt_as_proposals=True,
                  **kwargs):
         super(RandomSampler, self).__init__(num, pos_fraction, neg_pos_ub,
-                                            add_gt_as_proposals, **kwargs)
-        self.num = num
-        self.pos_fraction = pos_fraction
-        self.neg_pos_ub = neg_pos_ub
-        self.add_gt_as_proposals = add_gt_as_proposals
+                                            add_gt_as_proposals)
 
     @staticmethod
     def random_choice(gallery, num):