To run the example as a script main.py
, please
-
download video from here and specify video location with argument
--path /path/to/video
(as specified in example readme) -
download additional model checkpoints, SPPE FastPose (AlphaPose) and ST-GCN, and place them in the same folder where you put models for object detectors (e.g. pretrained YOLOX model)
The two model checkpoints should have name
fast_res50_256x192.pth
(SPPE FastPose) andtsstg-model.pth
(ST-GCN).When running the example as the script, specify where the above two models are located with argument
--model
. The command should look like:python VQPy/examples/fall_detection/main.py --path /path/to/video --save_folder /path/to/output/folder --model_dir /path/to/pose/detection/model/folder