Skip to content

Instantly share code, notes, and snippets.

View FrancescoSaverioZuppichini's full-sized avatar
🧸
Focusing

Francesco Saverio Zuppichini FrancescoSaverioZuppichini

🧸
Focusing
1k followers · 165 following
  • roboflow
  • Lugano, Switzerland
View GitHub Profile
@FrancescoSaverioZuppichini
FrancescoSaverioZuppichini / non-max-suppression-in-pytorch-20.py
Created October 26, 2022 07:57
Test
(bboxes ...)[order[i + 1:]]
@FrancescoSaverioZuppichini
FrancescoSaverioZuppichini / non-max-suppression-in-pytorch-19.py
Last active October 27, 2022 08:34
Test
(bboxes[...] * keep[...][...,None])
@FrancescoSaverioZuppichini
FrancescoSaverioZuppichini / non-max-suppression-in-pytorch-18.txt
Last active October 27, 2022 08:31
Test
iou = box_iou(bbox[None,...], (bboxes[order[i + 1:]]) * keep[i + 1:][...,None])
@FrancescoSaverioZuppichini
FrancescoSaverioZuppichini / non-max-suppression-in-pytorch-17.py
Last active October 27, 2022 08:37
Test
bbox = bboxes[order[i]]]
@FrancescoSaverioZuppichini
FrancescoSaverioZuppichini / non-max-suppression-in-pytorch-16.py
Created October 26, 2022 07:57
Test
if keep[i]:
@FrancescoSaverioZuppichini
FrancescoSaverioZuppichini / non-max-suppression-in-pytorch-15.py
Created October 26, 2022 07:57
Test
for i in indices:
...
@FrancescoSaverioZuppichini
FrancescoSaverioZuppichini / non-max-suppression-in-pytorch-14.py
Created October 26, 2022 07:57
Test
keep = torch.ones_like(indices, dtype=torch.bool)
@FrancescoSaverioZuppichini
FrancescoSaverioZuppichini / non-max-suppression-in-pytorch-13.py
Created October 26, 2022 07:57
Test
indices = torch.arange(bboxes.shape[0])
@FrancescoSaverioZuppichini
FrancescoSaverioZuppichini / non-max-suppression-in-pytorch-12.py
Created October 26, 2022 07:57
Test
order = scores.argsort()
@FrancescoSaverioZuppichini
FrancescoSaverioZuppichini / non-max-suppression-in-pytorch-11.py
Last active October 27, 2022 08:31
Test
from torchvision.ops.boxes import box_iou
def nms(bboxes: torch.Tensor, scores: torch.Tensor, iou_threshold: float) -> torch.Tensor:
order = torch.argsort(-scores)
indices = torch.arange(bboxes.shape[0])
keep = torch.ones_like(indices, dtype=torch.bool)
for i in indices:
if keep[i]:
bbox = bboxes[order[i]]
iou = box_iou(bbox[None,...],(bboxes[order[i + 1:]]) * keep[i + 1:][...,None])