Tensor.Art
Create

Lucky Strike Mix

CHECKPOINT
Reprint


Updated:

14K

All credit to kinshin007

Version 0.2 improves hand performance

The specific method is to replace all ATTN1 block information with AOM2-NSFW through the script,

The processed model is then blended with the model from the v0.1 pipeline,

But there was a problem of image degradation in the process, so I only mixed in about 50% of the hand correction.

The characteristic tradition of "Lucky Strike" is still preserved (laughs).

The original intention of this model is to draw slender and slender thighs,

This starting point is basically satisfactory, but the side effect is that I can't draw hands at all.

The name Lucky Strike Mix is ​​also derived from this, if there are enough times, there will always be a perfect picture

The following is the composition of the model

base model:

0.7(0.5(Gf_style2) + 0.5(mixedmixedmixedV2)) + 0.3(basil_mix)

replace OUT02,OUT04 with nightSkyYOZORAStyle_yozoraV1

--to get legs

replace IN00,IN01 with AbyssOrangeMix2_nsfw

--to get fingers

then reset all clips to get basic weight set

Although the directly generated pictures are hand-painted, I kept most of the real light and shadow, so it can be used with the real Lora to draw more realistic pictures

It should be noted that only standing legs are perfect (for me), other postures are still very error-prone

Version Detail

SD 1.5
All credit to kenshin007 Version 0.2 improves hand performance The specific method is to replace all ATTN1 block information with AOM2-NSFW through the script, The processed model is then blended with the model from the v0.1 pipeline, But there was a problem of image degradation in the process, so I only mixed in about 50% of the hand correction. The characteristic tradition of "Lucky Strike" is still preserved (laughs). The original intention of this model is to draw slender and slender thighs, This starting point is basically satisfactory, but the side effect is that I can't draw hands at all. The name Lucky Strike Mix is ​​also derived from this, if there are enough times, there will always be a perfect picture。 The following is the composition of the model base model: 0.7(0.5(Gf_style2) + 0.5(mixedmixedmixedV2)) + 0.3(basil_mix) replace OUT02,OUT04 with nightSkyYOZORAStyle_yozoraV --to get legs replace IN00,IN01 with AbyssOrangeMix2_nsfw --to get fingers then reset all clips to get basic weight setAlthough the directly generated pictures are hand-painted, I kept most of the real light and shadow, so it can be used with the real Lora to draw more realistic pictures It should be noted that only standing legs are perfect (for me), other postures are still very error-prone

Project Permissions

Model reprinted from : https://civitai.com/models/13034

Reprinted models are for communication and learning purposes only, not for commercial use. Original authors can contact us to transfer the models through our Discord channel --- #claim-models.

Comments

Related Posts

Describe the image you want to generate, then press Enter to send.