Works best with very simple prompts, no long list of negatives or refiner needed. Just use it right out of the box and have fun.
Example images are upscaled, I have added the original image in the comments on each image.
VAE is baked in. Use CLIP 1.
Description
FAQ
Comments (9)
Looks very nice but can you provide examples of initial txt2img?
Read the model description, read the image comments.
CarDos Anime is one of my favorite checkpoints, I'm looking forward to experimenting with this new XL version!
This isn't the Anime version sadly, He made two versions on 1.5 too, this is the photoreal one. Can't wait for the anime version which also was my favorite. We are still lacking a flat anime style LoRa for XL...
@Aerilyn235 I've been having success getting illustrated results with this and ZavyYumeXL by adding "photographic, photograph, realistic, realism, 35mm film, dslr" into the negative prompt.
It seems like SDXL is a lot more "determined" to make specific styles, so it takes a bit of creative encouragement to bump it onto something else.
Either way, I saw an XL version of YesMix was just released, which might interest you? Let me know if it scratches that anime itch!
@theharber What is the actual name? searching for YesMix only give 1.5 models.
my favorite XL model. blends so well with LoRA personas. great job.
all i see is black screen of death