👉 this is an illustrious-based modelmerge that combines: 3 trained base-models: Hassaku XL (Illustrious) + NoobAI-XL (NAI-XL) EPS + Obsession (Illustrious-XL) EPS; with 2 LORAs for better details: Detail Slider LoRA | Illustrious-XL + Detailer | Tool / Concept LoRA IllustriousXL; with illustriousXL (FP32 CLIP) both Clip-L and Clip-G by FellDude in one checkpoint for your pleasure💦... also: this checkpoint can be used in A1111, Forge and ComfyUI !!!
P.S.: add my suggested (in this model-card) resources at ~0.6667 strength for more artful output of your incredible image generations ;)
Description
👉 this version was improved with additional knowledge 🧠, and now you can easily generate beautiful images up to 1536px without any upscaling!
FAQ
Comments (14)
Ive been waiting for this day!, gotta try it!!.
IK, this model was made for YOU ;)
@OliviaRossi hshshs awwhh Tysm Oli <3, missed you. i wanted to try your flux models too, but my potato pc cant run flux hshs. one day tho, one day ill try them :3
@Razuryam <3
So there is no LoRa used for the preview images, and it's still this consistent? :0
yes, it was simply the same seed with changed some tokens (words in prompt)
@OliviaRossi Ah okay. Does it require some different denoise settings or a different KSampler?
@Joen98 sampler / scheduler combos for illustrious models have somewhat different "feel" of the output style - and you have to use the one that fits your taste - impossible to give recommendations here
🙏🙏🙏 Thank you so much Olivia!!! 🙏🙏🙏
TY, I'm already testing the update for this model ;)
i'll keep mine ,after see u !
<3 <3 <3
I'm amazed at how close this can hit to the Honkai Star Rail style without a LoRA!
yes, agree, but my experiments with LORAs give even broader style variations - so far I love this model very much ;)