wmcnally / kapao

KAPAO is an efficient single-stage human pose estimation model that detects keypoints and poses as objects and fuses the detections to predict human poses.
GNU General Public License v3.0
753 stars 103 forks source link

Where is the inference code in project #85

Open suijua opened 2 years ago

suijua commented 2 years ago

Hi! I just can not find the code of inference. specially, the code of "Keypoint Object Fusion (ϕ)" in the project, can you help me. thanks for you !

wmcnally commented 2 years ago

see post_process_batch in val.py

suijua commented 2 years ago

see post_process_batch in val.py

Ok, thank you

suijua commented 2 years ago

hi! thank you for reply. I has question about why didn't the code of "Keypoint Object Fusion (ϕ)" exist in the train.py file?