The first time I used mmaction2 for a demo inference, I was confused.

[nlxxWangYH]

The example of using the official demo is as follows

There are some questions of incomprehension

Why the Performing Human Detection for each frame progress is zero?
Why Performing SpatioTemporal Action Detection for each clip progress 27/28?
Whether the "problem" mentioned above is really a problem, and if so, what impact it will have, and how it should be solved.
I know my question is naïve, but as a newbie I am really confused, I hope to get answers from everyone, thank you very much！

[cir7]

I guess the progress is flushed by the following warning message, and has no effect on the result.