View Post Evaluation
Diffusion Models: From Noise to Images
32
The estimated attendance was 20 and the total was higher than this. I think that we underestimated how interested people would be in this event, as usually turnout for events in later weeks fall off dramatically. I think we can decrease this difference by having more faith in some of our events being interesting enough to gather a large number of people despite being later in the quarter.
199.34
The goal of this event was to give an introduction to the math behind diffusion and generative art models. It provided both a presentaton with theoretical slides as well as a notebook to see implementations in a hands-on setting.
I would consider the event to have been massively successful due to the extremely high turnout and engagement from the audience during the presentation. They were actively asking questions and following along well. Additionally most people stayed to the end of the event.
No complications arose during this event.
I would focus more on the hands-on learning and the notebook for the next event, and less on the presentation portion.
I would advertise the event with demos from the notebook, as I think they have the potential to be more interesting in this particular case of generative models than other forms of advertisement.