Visitar URL original
New Top-Down Pose Estimation Model - RTMPose by n-poulsen · Pull Request #2795 · DeepLabCut/DeepLabCut · GitHub
Skip to content

New Top-Down Pose Estimation Model - RTMPose #2795

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
n-poulsen merged 53 commits into from
Dec 20, 2024
Merged

New Top-Down Pose Estimation Model - RTMPose #2795

Changes from 1 commit
Commits
Show all changes
53 commits
Select commit Hold shift + click to select a range
3678c57
added rtmpose head
Oct 22, 2024
094ba6f
added CSPNext backbone
Oct 22, 2024
ebebb3b
add code to load pretrained CSPNext weights
Oct 22, 2024
ed0cbee
update number of backbone channels
Oct 22, 2024
6a66683
update weight init
Oct 22, 2024
34f1baf
added possibility for SequentialLR, tests, docs, default for RTMPose
Oct 24, 2024
f737e53
fix scheduler/optimizer config update
Oct 25, 2024
62e288f
fix tests
Oct 28, 2024
0b49323
made top down crop work on non-square images, in numpy to make faster :)
Oct 30, 2024
6ed1ab3
added example to run top-down video analysis without the detector
Oct 30, 2024
8aab808
improved inits
Oct 30, 2024
f13f88e
edit rtmpose_x for default square bbox
Oct 30, 2024
5add563
add random bbox transform; fix top_down_crop
Oct 31, 2024
a8470d2
add missing transform method
Oct 31, 2024
27679eb
fix transform - label must be int
Oct 31, 2024
5ae8bea
only apply top-down crop on non-empty image
Oct 31, 2024
57da127
bug fix: collating scales
Oct 31, 2024
49c6a4a
set image dtype when no bounding box
Oct 31, 2024
eebbfe5
fix simcc keypoint weights when -1 or 2
Oct 31, 2024
083dcba
update scheduler settings for rtmpose
Oct 31, 2024
5d2dc6c
fix rtmpose-m config
Oct 31, 2024
91e40eb
bug fix: bbox from keypoints
Nov 1, 2024
1697e12
update lr and switch to linear warmup
Nov 1, 2024
8e6ad27
bug fix - input size
Nov 1, 2024
3635698
added documentation
Nov 1, 2024
76b1414
added code to download backbone weights from huggingface
Nov 7, 2024
1bd0564
bug fix; added s config
Nov 8, 2024
5f9dd9c
fix LR to 5e-05 after cosine annealing LR
Nov 8, 2024
26d9227
init learning rate to 5e-3
Nov 8, 2024
fe88d4f
Merge branch 'pytorch_dlc' into niels/rtmpose
Nov 8, 2024
be15e9e
update default LR
Nov 8, 2024
d8d8581
Merged pytorch_dlc changes
Nov 8, 2024
4528fab
Merge branch 'pytorch_dlc' into niels/rtmpose
Nov 11, 2024
3ee1f4c
Merge branch 'pytorch_dlc' into niels/rtmpose
Nov 22, 2024
f336d2d
Merge branch 'pytorch_dlc' into niels/rtmpose
Nov 25, 2024
97970fe
fix rtmpose_s config
Nov 25, 2024
694b467
bug fix - target generation
Nov 25, 2024
b18c69b
bug fix: CSPNeXt config
Nov 28, 2024
e3e7513
update user guide
Nov 28, 2024
445ad4a
update user guide
Nov 28, 2024
6938ee4
Merge branch 'pytorch_dlc' into niels/rtmpose
Dec 3, 2024
2c3e6bd
Merge branch 'pytorch_dlc' into niels/rtmpose
Dec 4, 2024
9d9cd34
fix NaNs propagating to non-visible keypoints
Dec 4, 2024
9d07696
update default configs
Dec 4, 2024
e4d8699
address github comments
Dec 10, 2024
7158a52
Merge branch 'pytorch_dlc' into niels/rtmpose
Dec 10, 2024
ba536a2
fix import for top_down_crop
Dec 12, 2024
4328068
Merge branch 'pytorch_dlc' into niels/rtmpose
Dec 13, 2024
9e075e9
crop sampling after affine aug
Dec 13, 2024
d187038
RTMPose helper method
Dec 13, 2024
ed403eb
bug fix: pose inference runner
Dec 13, 2024
7f6b092
Merge branch 'pytorch_dlc' into niels/rtmpose
MMathisLab Dec 14, 2024
279adba
Merge branch 'pytorch_dlc' into niels/rtmpose
Dec 19, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
Prev Previous commit
Next Next commit
bug fix - target generation
  • Loading branch information
Niels Poulsen
Niels Poulsen committed Nov 25, 2024
commit 694b46738556dfb9c1d4f2d5c22a43e2efe07baa
2 changes: 1 addition & 1 deletion deeplabcut/pose_estimation_pytorch/models/target_generators/sim_cc.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -174,7 +174,7 @@ def _map_coordinates(
"""Mapping keypoint coordinates into SimCC space"""
keypoints_split = keypoints.copy()
# set non-visible keypoints to 0; deals with NaNs
keypoints_split[~keypoints_visible] = 0
keypoints_split[keypoints_visible <= 0] = 0
keypoints_split = np.around(keypoints_split * self.simcc_split_ratio)
keypoints_split = keypoints_split.astype(np.int64)
keypoint_weights = (keypoints_visible > 0).astype(keypoints_split.dtype)
Expand Down