JJJ · Sleight - Ver 1.0

JJJ · Sleight

LORA
Original


Updated:

JJJ · Sleight by JunJuneJunie on Tensor.Art
JJJ · Sleight by JunJuneJunie on Tensor.Art
JJJ · Sleight by JunJuneJunie on Tensor.Art
JJJ · Sleight by JunJuneJunie on Tensor.Art
JJJ · Sleight by JunJuneJunie on Tensor.Art
JJJ · Sleight by JunJuneJunie on Tensor.Art
JJJ · Sleight by JunJuneJunie on Tensor.Art
JJJ · Sleight by JunJuneJunie on Tensor.Art
JJJ · Sleight by JunJuneJunie on Tensor.Art
JJJ · Sleight by JunJuneJunie on Tensor.Art
JJJ · Sleight by JunJuneJunie on Tensor.Art
JJJ · Sleight by JunJuneJunie on Tensor.Art
JJJ · Sleight by JunJuneJunie on Tensor.Art
JJJ · Sleight by JunJuneJunie on Tensor.Art
JJJ · Sleight by JunJuneJunie on Tensor.Art
JJJ · Sleight by JunJuneJunie on Tensor.Art
JJJ · Sleight by JunJuneJunie on Tensor.Art
JJJ · Sleight by JunJuneJunie on Tensor.Art
JJJ · Sleight by JunJuneJunie on Tensor.Art
JJJ · Sleight by JunJuneJunie on Tensor.Art
JJJ · Sleight by JunJuneJunie on Tensor.Art
JJJ · Sleight by JunJuneJunie on Tensor.Art
JJJ · Sleight by JunJuneJunie on Tensor.Art
JJJ · Sleight by JunJuneJunie on Tensor.Art

Sleight is the magician I hand a prompt to when I already like the damn prompt.

She usually works with what's there before she starts redecorating the whole image. The clothes, the scene, the mood, that old prompt you still keep coming back to. She's more likely to let those things do their job without immediately turning everything into her personal art project.

Usually. She can still look at a plain sleeveless top and decide it needs a hood. Girl, I don't know either.

She's the quieter one of this trio. Quieter. Keep that little "-er" in mind.

What she tends to do

  • Keeps more of a prompt's existing rendering direction, especially when the medium is left open. In my old-prompt retests, she was less eager than Illusion and Morphing to redraw every face in her own illustrated style.

  • Can move between softer, photo-like detail and painted character work when the prompt calls for it.

  • Handles more than portraits: ordinary animals, dry materials, food, layered paper, small environments and mechanical objects all have useful results in the reviewed batches.

  • Can keep ordinary bodies, mature faces and expressions such as fatigue or open laughter. She isn't limited to one glamorous face.

Where she still gets ideas

Exact ages can drift, and some everyday clothes pick up extra design choices. A readable pair of hands doesn't guarantee a fiddly action is right. Subtle human-animal traits, chest binding and other precise anatomy or clothing instructions still need checking.

"Quieter" is a comparison within this trio, not a promise that every detail will obey.

How to use her

The main comparison galleries use strength 1.0. You don't have to. If you want a lighter touch, start lower and adjust from there. I haven't pinned down one perfect lower strength for every prompt.

She's a useful first try when you're bringing over an older prompt and want to see how much of its original feel carries through. If you want a more opinionated illustrated take, try JJJ · Illusion. If you want her sister to turn a few things way up, try JJJ · Morphing.

Settings

Tested base: rRealisMAX Krea-2 V1.5-aio

Sampler / scheduler: Euler / beta

Steps: 8-10; the main comparison galleries use 10

CFG: 1 Strength Main reviewed batches: 1.0; lower strengths are worth trying for everyday use

Trigger: The reviewed prompts did not include jjjmosaic; you can start without it

Negative: Left empty in the reference runs

VAE: Automatic

Reference size: 768 x 1152

The 8-step starting point comes from my additional use, while the structured comparison batches use 10. Lower-strength advice is a place to start experimenting, not a measured sweet spot. Other bases and workflows may behave differently.

About this version

This is the E4 checkpoint from the same mixed-dataset training as JJJ · Illusion (E7) and JJJ · Morphing (E10). They are being released for different uses rather than as one mandatory upgrade path.

The structured review covers 87 prompts for this checkpoint, across characters, groups, animals, objects, environments and several medium directions. The comparisons use different seeds, so they describe the images I got rather than isolate every difference to the LoRA alone.

Version Detail

Trained by Tensor
KREA_2
2
4

Project Permissions

Reprinting is strictly prohibited

    Use Permissions

  • Use in TENSOR Online

  • As a online training base model on TENSOR

  • Use without crediting me

  • Share merges of this model

  • Use different permissions on merges

    Commercial Use

  • Sell generated contents

  • Use on generation services

  • Sell this model or merges

Related Posts