How to use an existing pretrained ML model with examples

Job ID: 32043698

Budget: €30 – €250 EUR

I'm trying to find "epic moments" in live content through deep learning, and I found this article (https://epjdatascience.springeropen.com/articles/10.1140/epjds/s13688-021-00295-6) which makes research that it's exactly what I want to achieve.

In the supplementary material (codes and embeddings), he provides the necessary code and models to make it work (full content in its GitHub https://github.com/dscig/twitch-highlight-detection).

I'm not an ML developer, so I don't know how to use the model to get the expected output, and I would like to know how to use this model so I can pass some input data and get the output which indicates if it is an epic moment or not. I'm used to typical functions f(x) which provides an output y (i.e. f(x) = y), and I really don't know how to pass some arguments to that trained model to get the output which indicates if it is an epic moment or not.

I would like to someone explain to me how this trained model works, and how can I use it so given the input of the model (chat messages, video frames, etc) it returns me if the moment is epic or not (with examples).
Related categories: Python Machine Learning (ML) Data Science Pytorch