Unverified 提交 8035b616 authored 作者: Glenn Jocher's avatar Glenn Jocher 提交者: GitHub

Update objectness IoU sort (#3786)

上级 157aa2f8
......@@ -89,6 +89,7 @@ class ComputeLoss:
# Compute losses
def __init__(self, model, autobalance=False):
super(ComputeLoss, self).__init__()
self.sort_obj_iou = False
device = next(model.parameters()).device # get model device
h = model.hyp # hyperparameters
......@@ -134,8 +135,9 @@ class ComputeLoss:
# Objectness
score_iou = iou.detach().clamp(0).type(tobj.dtype)
sort_id = torch.argsort(score_iou)
b, a, gj, gi, score_iou = b[sort_id], a[sort_id], gj[sort_id], gi[sort_id], score_iou[sort_id]
if self.sort_obj_iou:
sort_id = torch.argsort(score_iou)
b, a, gj, gi, score_iou = b[sort_id], a[sort_id], gj[sort_id], gi[sort_id], score_iou[sort_id]
tobj[b, a, gj, gi] = (1.0 - self.gr) + self.gr * score_iou # iou ratio
# Classification
......
Markdown 格式
0%
您添加了 0 到此讨论。请谨慎行事。
请先完成此评论的编辑!
注册 或者 后发表评论