I would like to measure the HALU consumption of my prompt chain.

Suggestion:

There is a sort-of "stopwatch" in the workspace that I can start and stop. It measures all the consumed HALU in the meantime and breaks them down for me so that I can see (for example):

- GPT4: 2.5 HALU

- Standard Runner: 0.4 HALU

- DallE: 0.5 HALU

etc.

I should track all consumption events that are triggered by my actions including consumptions from down-the-line function calls etc.