[Feature] Video analysis

[tekntrash]

Feature request

The models included seem to be focused only on text and image. In robotics we use models able to interact between video and commands, so we've wonder if those are present in this repo

Motivation / references

robotics foundation models (RFMs) integrate video perception, natural language commands, and movement control into a unified AI framework

Your contribution

having it this repo would only make sense, as it seems to be a unified way of dealing with models of different kinds