Context

Currently running the Who App on localhost and only did a partial crawl and there are 6k+ users:

That's already enough to require some scrolling when viewing the "wall of faces" ... 📜

This will only increase when we run the full crawl in prod ...
I expect around 35k unique users starring across all @dwyl repos 👤 ⭐

Problem

The problem with this is we want to display a "Wall of Faces"; 😍
not a wall of users who couldn't be bothered to change their profile picture. 🙃

It's especially visible in the large block of "space invaders" avatar images:

But if you look carefully at the rest of the wall above, ⬆
there are many other non-human-face avatar images. 😕

Task

• Get the image-classifier project running on your localhost https://github.com/dwyl/image-classifier T1d
  e.g. on MBP "16 #38

• Figure out if we can feed an avatar image into the existing model and have it confirm if it includes a human face.

• Deploy the project to Hetzner on a $3.99/month (4GB RAM server) hetzner learn-devops#97 with an API key to protect it from abuse.

priority-2 because this is a super nice-to-have; not costing us any cash, but would be awesome to have.
should only be worked on after we've deployed the who app to prod