How can I inference it on random video clip?

Hi Team,

The code provided is indeed fancily built on Hydra config structure for better changing in the different parameters for all the models provided. I couldn't find a solution for the inference code from the repo?