I haven't revealed -too- much about the new DeOldify model I'm working on yet. But two things it does a lot better with are crowds and buildings. Behold!https://twitter.com/citnaj/status/1191060069854543873 …
You can add location information to your Tweets, such as your city or precise location, from the web and via third-party applications. You always have the option to delete your Tweet location history. Learn more
I'd love to read a mind dump of how you go about improving / tweaking your models. With NNs, it feels like there's an endless frontier of possible paths and it's really hard to know when a particular approach might work if you put more effort and when it's a dead end.
Endless experimentation/heating up the house coupled with constantly questioning my previous conclusions/sanity lol
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.