Issue with local run on a custom video input

Hi,

Can someone from the team assist in my endeavor to be able to use SlowFast using AVA dataset to be able to predict 17 out of 80 classlabels for person to person interaction for my own video clips

I am trying to run run_net.py with below parameters
--cfg C:\Self\ML\code\SlowFast-master\demo\AVA\SLOWFAST_32x2_R101_50_50.yaml TRAIN.ENABLE False TEST.ENABLE False DEMO.ENABLE True NUM_GPUS 0

I just get a gray box on screen, my webcam is on and iterations continue. There is a white bounding box on gray popup, but nothing else. Attached is output snapshot and the yaml file as txt extension for reference


[SLOWFAST_32x2_R101_50_50.yaml.txt](https://github.com/facebookresearch/SlowFast/files/6385718/SLOWFAST_32x2_R101_50_50.yaml.txt)