I used a large number of Flux-generated images to create this upgrade file. It makes generating Asian faces easier, reduces the rate of distorted hands, and delivers sharper focus.
this is for all friends who, like me, love realistic Illustrious.
Try it online: Tensor | Seaart | Liblib
Hope you all have fun with it!
Parameters:Steps: 20-40, CFG scale: 4-7
Samplers: Euler a/dpmpp_sde/dpmpp2m_sde
Schedulers: karras/sgm_uniform/exponential/beat
VAE is pre-integrated.
My go-to settings: 25 steps, CFG 6, dpmpp_sde sampler, beat scheduler, 832x1536 resolution.
For the record, I’ve always trained and updated this checkpoint independently, without merging it with any other checkpoints. Please note: reselling this model or selling merged versions of it is strictly prohibited.
The base model used in the merging process is xRicaMix_ILXL_v5.2. For facial restoration, please switch to XLVAE.
Description
use size larger than1024x1024 for the original dimensions.
My go-to settings: 25 steps, CFG 6, dpmpp_sde sampler, beat scheduler, 832x1536 resolution.
FAQ
Comments (20)
Thank you for making such a great model
Thank you, bro
Your big model is so beautiful, I like it very much. I hope you can make a big model in the style of Pony. I have many Lora that are Pony models, which are too sad to be displayed.
Illustrious is compatible with some pony loras, so you can try using them.
Thank you for the beautiful women!
But it's difficult to display multiple human.
This is just a realistic type checkpoint that takes into account Danbooru tags. It is not created just for NSFW.
Bro it's really a nice model, thank you so much for your working.
Type: Checkpoint Trained (The merge after training is still classified as training. I remember someone put forward this point of view before.)
when i upscale without using vae (vae is from model), why does it produce green artifacts, what vae do i need???
VAE is my own creation and has not been published. To get better skin color, please replace it with XL VAE when you need upscaler.
The next version will be coming soon. I have replaced it with a normal VAE. There will be no green artifacts when training Lora, ADetailer and upscaler. I should release it in a week.
@Ray3780 thanks so much, your model are really the best!!!
@Ray3780 May I ask when the checkpoints will be updated?🫠
@nanayoda Because of my mistake, it was placed on another page.
@Ray3780 OK，I get it!
我尝试了很多次，下载总是莫名其妙的中断。
All images are using <lora:None:1.0> in promt. What real name of this lora, and where to download it?
Actually, there's no LoRA involved here. I'm using the Efficiency Nodes for ComfyUI, which automatically insert the <lora:None:1.0> tag. If your results aren't up to par, it's likely because you aren't using the dpmpp_sde + Beta sampler.
Which one lora that you merge into the checkpoint that give the auroral blushing skin like that?
Details
Files
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.