Tip: Since it is an art-style LoRA, it might not do too well with faces while performing standalone creation. It is recommended to put in a pre-created image as reference (use PixAI, prodia, LensGo, Dreamlike Art, Adobe Firefly, Dall-E, MS Bing image creator, etc. to get the reference image, or use one you like). Now, try it at strength of 55% and change reference image strength according to your needs and results produced.
This LoRA is trained on Realistic Vision v5.1 using high-res screengrabs, wallpapers and iconic pictures from Blade Runner 2049. It aims to recreate the ambiance, visual aesthetic, cinematography and architecture from the film.
This should work well with any realistic models.
Give this one a try and let me know how it turns out.
