Warcraft LoRA Overview

This LoRA was trained on the artworks of Eva Widermann, a great artist and known well for her contributions to Blizzard, Wizards of the Coast, Disney, and Paizo. Her art is most recognizable from the card art she provided for Hearthstone, but she has been an illustrator for a long time and her art has inspired countless TTRPG and RPG enthusiasts alike.

LoRA - What is it?

Well, it is an acronym for Low-rank Adaptation, but that doesn't really say what it is. LoRA is a method used for quick fine-tuning of diffusion models. What you end up with after training is a much smaller 'model' that works with your other models. It is used somewhat like a hypernetwork, but currently the tool most people use with LoRA models is Automatic1111.

Recommended Settings

• LoRA strength - between 0.4 and 0.75 give the best results. I use 0.6 mostly, but you can play around with the strength and tweak to your liking

• I used Elldreth's Vivid Mix V2.0 to train it and get good results with that model, but it also works really well with TheAlly's Mix II: Churned and Lykon's DreamShaper 3.3. You might also find it works well with other models, so feel free to share!

• Highres.fix is great with this LoRA but you can get good images without it

Please feel free to leave your feedback and share your creations! Enjoy!

Special Thanks to Lykon for helping me get my LoRA settings tweaked correctly and for the advice that has greatly helped me out. Also thanks to TheAlly (The Prompt Queen) for always helping me and allowing me to use some of her prompts for preview. She has a lot to offer and if interested you can find her Patreon here: https://www.patreon.com/theally