stylegan2-pytorch VS og-image

GANs on the other hand uses a generator and discriminator model, where the generator learns to generate images that are "good enough" to fool the discriminator. It may make it easier to make discriminators that are better at picking out visual artefacts that stand out for humans, (it was easy to pick out "fake humans" from the early versions of thispersondoesnotexist.com by looking at their eyes and teeth, they tend to be pointing in different directions, or their teeth are too irregular before the discriminator got better at picking those out) but they can suffer from mode collapse, and generate images with very little variety.

Generative networks like StyleGAN

It's interesting that AI generated pictures of people can get most things right but seemingly always struggles with teeth & fingers. If you check out thispersondoesnotexist.com the teeth are almost always goofy looking...

Teeth as well, check out thispersondoesnotexist.com you'll come across a lot of convincing looking people with odd teeth

https://thispersondoesnotexist.com/ has been around for 4 years now...

I remember there was https://thispersondoesnotexist.com/ Did it also work on StableDiffusion beneath?

Another great place to see AI generated faces: https://thispersondoesnotexist.com/

Check out https://thispersondoesnotexist.com/

Now here's the thing since these are part of the meta tags, and The image shown in the tags are ... well! images, making them dynamic is a bit tricky. generating images isn't a menial task that just every web server can do ... it's resource intensive, time-consuming and all the other nightmares. and making it part of your dynamic web app is another ball game altogether. So what's new? well, @vercel/og is, this library lets you build image content from HTML/React that too insanely fast, like really fast . and you can pair this with any edge network like AWS Lambda on the edge, Cloudflare workers, etc to make it even faster.

Thanks! Not to discourage anyone from using Vercel services if they’re interested (y’all are doing great stuff on this and lots of other fronts), but! I’d be remiss if I didn’t point out the unpublished-but-MIT-licensed repo I could find.

For anyone else who’s curious it looks like the pertinent source[1] can be self-hosted. I’m on mobile so I’m gonna limit my peering into the source, but it appears to wrap Chromium to do the PNG generation. Quite a bit different from my solution (which takes SVG-producing JSX and anything producing CSS, and renders to PNG with Sharp).

I’m curious how much overhead using Chromium adds, and whether alternatives like Sharp were considered.

1: https://github.com/vercel/og-image

It’s not embedding. When you paste that image url a serverless function running a headless browser requests say a next app which loads up a page based on the params in the url. Then we take a snapshot and image data is sent back as response to you hence the image appears. You can check out this: with NextJS and puppeteer

But what if the content of the page is not visual at all, and no photo really suits the contents of the page? And what if we have a lot of these pages, with dynamically generated content? Luckily we are not limited to photography, we can use text! The prime example of this is how GitHub generates previews for issue URLs: they generate an image containing enough info such that users know what they will click on. These images are generated on-demand, as storing an image for every GitHub pull request or issue would be unfeasable. Another example is Vercel, which even open sourced their Open Graph Image as a Service that they occasionally use on blog posts or announcements.

Hi, it is basically a embedded, better and easier (API-wise) version of https://github.com/vercel/og-image

Vercel has a really good simple generator if you don’t want to spend time on graphic design: https://og-image.vercel.app

I ended up choosing Puppeteer and capture the web page, which is pretty far from what I initially wanted. Puppeteer needs more time and memory than I prefer. I can generate an image, download it, and then use it as a cover image. But, I want to create an API (with Next.js API routes) that functions as Open Graph Image as a Service. I later found out that Vercel already made this 🥲, still, I want to create my own.

That sounds a great side project, since some developers might prefer a static svg dynamically created via url like boringavatar.app/[username here].

There's some prior work on this: e.g. https://github.com/lfades/static-tweet and https://github.com/vercel/og-image.

If you are a developer, you've probably seen the Open Graph images (part of Open Graph Protocol) generated by popular dev related websites like DEV.to or even Vercel's Open Graph Image as a Service. Both examples are using an approach to render image related to the content, so it contains some standard layout background, an image in it that is related to the content (vercel's logo or author's avatar), headline or title of the article, and a description.