This is my own rendition of my favorite SD 1.5 model ported to SDXL. This model was trained on 5000 images generated using a custom LoRa combined with ReV Animated! Credit goes to s6yx for his creation of Rev Animated
Style: This model has the ability to create a variety of styles but focuses on 2.5D like images.
Kind of generations:
It works best on these resolution dimensions:
This model has some difficulty with hands when describing complex prompts
It wasn't trained on any realistic images, so I cannot confirm if it has that capacity.
Not trained on any NSFW work
Steps 30-40 seem to work good
CFG of 4-10 was experimented with, it tends to work better at CFG 6-9
Hires fix works well for some images, try putting the denoise at 0.55 and upscale at 1.4
Please experiment with your prompting, this was trained using no captions so you will find that sometimes prompting can cause mutations to occur in the images, This is generally the price we pay for having incredible composition and stylization in our models. There is so much you can create in this model, I believe it maintains the versatility that SDXL 1.0 has while allowing for greater color scheme, details, and composition while lending toward a semi-realistic 2.5D style.
Do not sell this model on any website without permissions from creator (me)
Credit me if you use my model in your own merges
I do not authorize this model to be used on generative services