XL version of my previous standard lora. A few changes in training data but should be very similar.
Used animagine base - didn't realise this made such a big difference in results so I'm now making an illustrious version (sadly pony has been removed as an option)
