I trained a model on Mostly the 19th century paintings you would see in a Art History Class. And a few older ones.
Description
FAQ
Comments (6)
Luv this!
Wasn't 1.5 already trained on most 19th century artists? Was there any need for more training?
A lot of subjects aren't rained enough. Yesterday I saw the example of Tintin - without an embedding, it doesn't even look like Tintin, but with an embedding, it's much better.
In 1.5 by itself you have to type specific artists and results vary, in here you get a more averaged result. I also edited out all the Putti Cherubs from the source. You get some unwanted results often from typing straight Bouguereau.
sometimes having specialized sets trained more explicitly on certain styles rather than a big dataset that's more of a "general" set can really aid with trying to generate something more specific and refined :)
Looks great! Please upload a safetensor.
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.