Name		Name	Last commit message	Last commit date
Latest commit History 277 Commits
data		data
models		models
utility		utility
.gitignore		.gitignore
README.md		README.md
args.py		args.py
debug.py		debug.py
ntu.ipynb		ntu.ipynb
optimizer.py		optimizer.py
requirements.txt		requirements.txt
samples_with_missing_skeletons.txt		samples_with_missing_skeletons.txt
samples_with_missing_skeletons_120.txt		samples_with_missing_skeletons_120.txt
train.ipynb		train.ipynb
train.py		train.py
train2s.py		train2s.py
train3.py		train3.py
trainer.py		trainer.py
trainer2.py		trainer2.py
vat.py		vat.py
vat_train.py		vat_train.py
visualization.ipynb		visualization.ipynb
visualize.py		visualize.py
wavelet.ipynb		wavelet.ipynb

Repository files navigation

APBGCN

Attention-Based Parts-oriented Graph Convolution Networks

download ntu rgb+d 60 action recognition from skeleton from http://rose1.ntu.edu.sg/datasets/actionRecognition.asp

uzip data as the following file structure: APBGCN/nturgb+d_skeletons/raw/.*skeleton

change default dataset_root in args.py : dataset_root=osp.join(os.getcwd(),"nturgb+d_skeletons"),