conda create --name openmmlab python=3.8 -y
conda activate openmmlab
conda install pytorch torchvision -c pytorch
pip install -U openmim
mim install mmengine
mim install "mmcv>=2.0.1"
mim install "mmdet>=3.1.0"
cd mmpose
pip install -r requirements.txt
pip install -v -e .
mim download mmpose --config td-hm_hrnet-w48_8xb32-210e_coco-256x192 --dest .
cd ..
cd LSTMAT
pip install -r requirements.txt
wget https://huggingface.co/allasobi110/LSTMAT_AI4L/resolve/main/LSTM_Attention_128HUs_f1.h5?download=true
-
Root_Folder: |____ LSTMAT |____ mmpose |____ Data_Round1/Data_R1 |_____ barbell biceps curl | |____tên video_1.mp4 | |____…………… | |____tên video_n.mp4 |_____ bench press | |____tên video_1.mp4 | |____…………… | |____tên video_n.mp4 |....................................... |_____ tricep Pushdown |____tên video_1.mp4 |____…………… |____tên video_n.mp4
chmod u+x run.sh
./run.sh
- The result is an image in folder LSTMAT named PIXCEL_SPACE_CORR_F1_RESULT-AAAAAAAAAAA.png