Skip to content

Commit 1dada9c

Browse files
Ben-LouisBen-Louis
authored
Merge e63cd04 into ce8da8f 
2 parents ce8da8f + e63cd04 commit 1dada9c

File tree

7 files changed

+27
-27
lines changed

7 files changed

+27
-27
lines changed

‎configs/body_2d_keypoint/yoloxpose/coco/yoloxpose_coco.md‎

Lines changed: 4 additions & 4 deletions
Original file line numberDiff line numberDiff line change
@@ -53,7 +53,7 @@ Results on COCO val2017 with detector having human AP of 56.4 on COCO val2017 da
5353

5454
| Arch | Input Size | AP | AP<sup>50</sup> | AP<sup>75</sup> | AR | AR<sup>50</sup> | ckpt | log |
5555
| :-------------------------------------------- | :--------: | :---: | :-------------: | :-------------: | :---: | :-------------: | :-------------------------------------------: | :-------------------------------------------: |
56-
| [yoloxpose_tiny](/configs/body_2d_keypoint/yoloxpose/coco/yoloxpose_tiny_4xb64-300e_coco-416.py) | 416x416 | 0.527 | 0.794 | 0.557 | 0.577 | 0.843 | [ckpt](https://download.openmmlab.com/mmpose/v1/body_2d_keypoint/yolox_pose/yoloxpose_tiny_4xb64-300e_coco-416-76eb44ca_20230829.pth) | [log](https://download.openmmlab.com/mmpose/v1/body_2d_keypoint/yolox_pose/yoloxpose_tiny_4xb64-300e_coco-416-20230829.json) |
57-
| [yoloxpose_s](/configs/body_2d_keypoint/yoloxpose/coco/yoloxpose_s_8xb32-300e_coco-640.py) | 640x640 | 0.642 | 0.873 | 0.702 | 0.688 | 0.912 | [ckpt](https://download.openmmlab.com/mmpose/v1/body_2d_keypoint/yolox_pose/yoloxpose_s_8xb32-300e_coco-640-56c79c1f_20230829.pth) | [log](https://download.openmmlab.com/mmpose/v1/body_2d_keypoint/yolox_pose/yoloxpose_s_8xb32-300e_coco-640-20230829.json) |
58-
| [yoloxpose_m](/configs/body_2d_keypoint/yoloxpose/coco/yoloxpose_m_8xb32-300e_coco-640.py) | 640x640 | 0.697 | 0.903 | 0.766 | 0.739 | 0.933 | [ckpt](https://download.openmmlab.com/mmpose/v1/body_2d_keypoint/yolox_pose/yoloxpose_m_8xb32-300e_coco-640-84e9a538_20230829.pth) | [log](https://download.openmmlab.com/mmpose/v1/body_2d_keypoint/yolox_pose/yoloxpose_m_8xb32-300e_coco-640-20230829.json) |
59-
| [yoloxpose_l](/configs/body_2d_keypoint/yoloxpose/coco/yoloxpose_l_8xb32-300e_coco-640.py) | 640x640 | 0.714 | 0.906 | 0.785 | 0.756 | 0.934 | [ckpt](https://download.openmmlab.com/mmpose/v1/body_2d_keypoint/yolox_pose/yoloxpose_l_8xb32-300e_coco-640-de0f8dee_20230829.pth) | [log](https://download.openmmlab.com/mmpose/v1/body_2d_keypoint/yolox_pose/yoloxpose_l_8xb32-300e_coco-640-20230829.json) |
56+
| [yoloxpose_tiny](/configs/body_2d_keypoint/yoloxpose/coco/yoloxpose_tiny_4xb64-300e_coco-416.py) | 416x416 | 0.526 | 0.793 | 0.556 | 0.571 | 0.833 | [ckpt](https://download.openmmlab.com/mmpose/v1/body_2d_keypoint/yolox_pose/yoloxpose_tiny_4xb64-300e_coco-416-76eb44ca_20230829.pth) | [log](https://download.openmmlab.com/mmpose/v1/body_2d_keypoint/yolox_pose/yoloxpose_tiny_4xb64-300e_coco-416-20230829.json) |
57+
| [yoloxpose_s](/configs/body_2d_keypoint/yoloxpose/coco/yoloxpose_s_8xb32-300e_coco-640.py) | 640x640 | 0.641 | 0.872 | 0.702 | 0.682 | 0.902 | [ckpt](https://download.openmmlab.com/mmpose/v1/body_2d_keypoint/yolox_pose/yoloxpose_s_8xb32-300e_coco-640-56c79c1f_20230829.pth) | [log](https://download.openmmlab.com/mmpose/v1/body_2d_keypoint/yolox_pose/yoloxpose_s_8xb32-300e_coco-640-20230829.json) |
58+
| [yoloxpose_m](/configs/body_2d_keypoint/yoloxpose/coco/yoloxpose_m_8xb32-300e_coco-640.py) | 640x640 | 0.695 | 0.899 | 0.766 | 0.733 | 0.926 | [ckpt](https://download.openmmlab.com/mmpose/v1/body_2d_keypoint/yolox_pose/yoloxpose_m_8xb32-300e_coco-640-84e9a538_20230829.pth) | [log](https://download.openmmlab.com/mmpose/v1/body_2d_keypoint/yolox_pose/yoloxpose_m_8xb32-300e_coco-640-20230829.json) |
59+
| [yoloxpose_l](/configs/body_2d_keypoint/yoloxpose/coco/yoloxpose_l_8xb32-300e_coco-640.py) | 640x640 | 0.712 | 0.901 | 0.782 | 0.749 | 0.926 | [ckpt](https://download.openmmlab.com/mmpose/v1/body_2d_keypoint/yolox_pose/yoloxpose_l_8xb32-300e_coco-640-de0f8dee_20230829.pth) | [log](https://download.openmmlab.com/mmpose/v1/body_2d_keypoint/yolox_pose/yoloxpose_l_8xb32-300e_coco-640-20230829.json) |

‎configs/body_2d_keypoint/yoloxpose/coco/yoloxpose_coco.yml‎

Lines changed: 18 additions & 18 deletions
Original file line numberDiff line numberDiff line change
@@ -15,11 +15,11 @@ Models:
1515
Results:
1616
- Dataset: COCO
1717
Metrics:
18-
AP: 0.527
19-
[email protected]: 0.794
20-
[email protected]: 0.557
21-
AR: 0.577
22-
[email protected]: 0.843
18+
AP: 0.526
19+
[email protected]: 0.793
20+
[email protected]: 0.556
21+
AR: 0.571
22+
[email protected]: 0.833
2323
Task: Body 2D Keypoint
2424
Weights: https://download.openmmlab.com/mmpose/v1/body_2d_keypoint/yolox_pose/yoloxpose_tiny_4xb64-300e_coco-416-76eb44ca_20230829.pth
2525
- Config: configs/body_2d_keypoint/yoloxpose/coco/yoloxpose_s_8xb32-300e_coco-640.py
@@ -31,11 +31,11 @@ Models:
3131
Results:
3232
- Dataset: COCO
3333
Metrics:
34-
AP: 0.642
35-
[email protected]: 0.873
34+
AP: 0.641
35+
[email protected]: 0.872
3636
[email protected]: 0.702
37-
AR: 0.688
38-
[email protected]: 0.912
37+
AR: 0.682
38+
[email protected]: 0.902
3939
Task: Body 2D Keypoint
4040
Weights: https://download.openmmlab.com/mmpose/v1/body_2d_keypoint/yolox_pose/yoloxpose_s_8xb32-300e_coco-640-56c79c1f_20230829.pth
4141
- Config: configs/body_2d_keypoint/yoloxpose/coco/yoloxpose_m_8xb32-300e_coco-640.py
@@ -47,11 +47,11 @@ Models:
4747
Results:
4848
- Dataset: COCO
4949
Metrics:
50-
AP: 0.697
51-
[email protected]: 0.903
50+
AP: 0.695
51+
[email protected]: 0.899
5252
[email protected]: 0.766
53-
AR: 0.739
54-
[email protected]: 0.933
53+
AR: 0.733
54+
[email protected]: 0.926
5555
Task: Body 2D Keypoint
5656
Weights: https://download.openmmlab.com/mmpose/v1/body_2d_keypoint/yolox_pose/yoloxpose_m_8xb32-300e_coco-640-84e9a538_20230829.pth
5757
- Config: configs/body_2d_keypoint/yoloxpose/coco/yoloxpose_l_8xb32-300e_coco-640.py
@@ -63,10 +63,10 @@ Models:
6363
Results:
6464
- Dataset: COCO
6565
Metrics:
66-
AP: 0.714
67-
[email protected]: 0.906
68-
[email protected]: 0.785
69-
AR: 0.756
70-
[email protected]: 0.934
66+
AP: 0.712
67+
[email protected]: 0.901
68+
[email protected]: 0.782
69+
AR: 0.749
70+
[email protected]: 0.926
7171
Task: Body 2D Keypoint
7272
Weights: https://download.openmmlab.com/mmpose/v1/body_2d_keypoint/yolox_pose/yoloxpose_l_8xb32-300e_coco-640-de0f8dee_20230829.pth

‎configs/body_2d_keypoint/yoloxpose/coco/yoloxpose_l_8xb32-300e_coco-640.py‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -1,4 +1,4 @@
1-
_base_ = './yolopose_s_8xb32-300e_coco-640.py'
1+
_base_ = './yoloxpose_s_8xb32-300e_coco-640.py'
22

33
widen_factor = 1
44
deepen_factor = 1

‎configs/body_2d_keypoint/yoloxpose/coco/yoloxpose_m_8xb32-300e_coco-640.py‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -1,4 +1,4 @@
1-
_base_ = './yolopose_s_8xb32-300e_coco-640.py'
1+
_base_ = './yoloxpose_s_8xb32-300e_coco-640.py'
22

33
widen_factor = 0.75
44
deepen_factor = 0.67

‎configs/body_2d_keypoint/yoloxpose/coco/yoloxpose_s_8xb32-300e_coco-640.py‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -132,7 +132,7 @@
132132
loss_bbox_aux=dict(type='L1Loss', reduction='sum', loss_weight=1.0),
133133
),
134134
test_cfg=dict(
135-
score_thr=0.001,
135+
score_thr=0.01,
136136
nms_thr=0.65,
137137
))
138138

‎configs/body_2d_keypoint/yoloxpose/coco/yoloxpose_tiny_4xb64-300e_coco-416.py‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -1,4 +1,4 @@
1-
_base_ = './yolopose_s_8xb32-300e_coco-640.py'
1+
_base_ = './yoloxpose_s_8xb32-300e_coco-640.py'
22

33
# model settings
44
widen_factor = 0.375

‎mmpose/evaluation/functional/nms.py‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -356,7 +356,7 @@ def nms_torch(bboxes: Tensor,
356356
idx, indices = indices[0], indices[1:]
357357
bbox = bboxes[idx]
358358
ious = iou_calculator(bbox, bboxes[indices])
359-
close_indices = torch.where(ious > threshold)[0]
359+
close_indices = torch.where(ious > threshold)[1]
360360
keep_indices = torch.ones_like(indices, dtype=torch.bool)
361361
keep_indices[close_indices] = 0
362362
groups.append(torch.cat((idx[None], indices[close_indices])))

0 commit comments

Comments
 (0)