The Lora is trained on MidJourney Images generated with my "Personalised Style" + Monochrome Vintage Art Deco/Sci Fi images.
The dataset is really specific and might not produce perfect results for all prompts.
Works without the trigger words ArsMJStyle, ArtDecoVintageFuture, but gives stronger results with them.
The following tags are not requred , but match well with the trained style and can create stronger "Vintage" effect:
Monochrome, Traditional Media, Sepia, Art Deco
Works great in the 0.3 - 0.8 range, depending on the Lora Stack you use.
Around 1 the Monochrome effect starts kicking in, even without prompting it.
I recommend starting ~1 and adjusting accordingly.
Description
Same recommended settings as the other versions.
The LoRA and the Trigger words have really strong effect on the image. If you don't use the Triggers and add color references you will still get the Aesthetic, but with much softer results.
FAQ
Comments (3)
#AwesomeSauce
I am really happy how the Flux Lora turned out 😁 The SDXL was good, Pony just refused to learn some concepts properly, but Flux is indeed Awesome 😁
Produces some great images- I got impatient and didn't read the model description fully, but for the poster-type images entering "art deco poster style" worked well. I think it's worth messing with the strength to see what different effects are produced, as suggested in the description
Details
Files
Available On (2 platforms)
Same model published on other platforms. May have additional downloads or version variants.