LoRA the Explorer service generates pictures in different styles for free

Enthusiasts have made the LoRA the Explorer website, which contains a collection of image generators in different styles, working on Stable Diffusion XL.

Everything works simply: select the appropriate style from the list on the left, enter a request in the box on the right and click Run. If you get an error about too high load, click Run a few more times until your place in the queue appears on the screen.

After a few minutes of waiting, the finished picture will appear. For example, here's a watercolor-styled generation of "a cat with a samurai sword drawn by Utagawa Kuniyoshi":

And this is a modern Russian house in the style of photographer William Eggleston:

Bonus: a frame from the manga about Ivan the Terrible.

Styles are not always applied correctly: for example, for the query "Cyril from The Witcher" in the style of JoJo's Bizarre Adventure really looks like Ciri, but nothing from the cult anime to her image added. But the simplified query "Cirill" was drawn by the neural network closer to the given style.

The result of the generation can be saved to the device or immediately published on the social network.

The site is as simple as possible and does not require registration. This also leads to a minus: processing a large number of requests takes time. At the time of testing, I had to spend about 5 minutes in the queue for each request. To speed up the process, you can open several tabs so that the service processes requests in parallel.

If you try, share your results in the comments!

LoRA the Explorer →