[HOW TO USE]

These are styles for SDXL. They're not trained on specific concepts or people, but a dataset of images with the same "vibe". Because of this, it can be difficult to effectively prompt the style you download, so make sure to look at the words the LORA is trained on in your UI of choice!

[KNOWN ISSUES]

Most of these Loras have a heavy bias towards women, especially Asian women. This is because of the nature of where I mostly get my dataset from, but I am working to diversify my datasets and update older models with a more diverse one too.

These models are known to cause some artifacts if kept at a weight of 1. I recommend for all models to stick between 0.5-0.7, but I'll add specific recommended weights for each model too.

[MODEL LIST]

Supernova - Futuristic sci-fi with a focus on portraits. Good at making vibrant, colourful photos with a spacey cyber feel.

90s NYC - A vintage, nostalgiac vibe with an emphasis on 90s fashion and streetwear.

ModernFable - Hard to describe. Ethereal and dreamy mixed with surreal horror elements (picture Japanese Idols in the backrooms). Called "Modern Fable" because it incorporates fantasy and surreal elements in normal "realistic" places.

Technicolor Heaven - Currently the easiest to use since I reformulated my tagging approach. Based on Old Hollywood Cinema, specifically films shot on Technicolor. Good at capturing classic glamour with beautiful vibrancy.