Closed jobyclash2 closed 7 months ago
Hello @jobyclash2
It looks like you've installed the latest version of mmaction2
, which includes breaking changes, and this open-source project wasn't updated to run on its latest version (hence why it's important to clone the exact versions of the sub-libs that this repo defines).
Anyway, why not try to use the docker image provided in the instructions? Would remove a lot of the hassle with the step-by-step installation.
Hi i was interested in this project and wanted to try it out in my windows machine. i am new to python and i followed your manual installation tutorial and reached endless chaos of issues. i think the read me is either outdated or wrong.
i cloned your project i installed python (Python 3.9.13) installed cuda and cuDNN from nvidea installed pytorch (pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu121)
i navigated to project main directory (where src etc... are located) i went below link and installed mmcv mmdet,mmaction etc... https://mmaction2.readthedocs.io/en/latest/get_started/installation.html#best-practices
done
now when i run the project like below
python src/demo/multimodial_demo.py video.mp4 demo.json --timestamps
i am getting errors
there is a requirements.txt inside requirements folder . am i supposed to install that as well? in documentation there is only mention about extra.txt
these are the errors i get do you know why this happen