This project uses yolo11 to train on HDMB51 data, which can be used for real-time monitoring, which can be used for source data such as video, URL, and camera.
First of all, you need to install the dependencies required by yolo11, which can be found on the official website of yolo11 for details.
Then you can run action_r2
Since the theory is trained on HDMB51 data, there are many actions that cannot be detected temporarily, so if you need to detect the actions you want, you can choose the corresponding dataset for training.
Go to the xclip folder
pip install -r requirement.txt
Then download your data, after your dataset is downloaded, you may need to change the format of your dataset to fit the code, please refer to the official website of Xclip for details.
Finally, you're ready to start training.