[LoRA] macro bear

模型简介：
This loRA is for macrophiles and bear lovers.
if you like my work, please support me and buy me a coffee : https://ko-fi.com/zeid_creator

:)

(Information for sdxl version at bottom)

You can now follow me on my new twitter account created especially for image creation, it's NSFW and all images were created with my lora :
Zeid (@Zeid_ai) / X (twitter.com)

↓ Yes, you can make other races than bears, see below. ↓

This lora is a concept lora and will continue to improve over time. The theme is, of course, all things macro and bear. Other races may be compatible with my lora, but keep in mind that it was train with only bears.

What this lora brings, the best is by example :
settings :
checkpoint : BB95 V12 / sampling : DPM++ 3M SDE Karras / 20 steps / Hires.fix 20 steps with latent upscaler
pos prompt : macro, bear, city, cityscape, furry, musclegut, shoulder armor, greaves, plackart, vambraces, thick arms, thick thighs, detailed fur

textual inversion : deformityV6 as a textual inversion

and now, let's put my LoRA with a weight of 0.8 (you can use 1 but it will be a little more tricky to get what you want) by default.

Carefull this lora is not compatible with VPRED for now. It's usable but the result can be very strange sometime.
If you want a nude character, you may have a problem with a belt appearing no matter what. To solve this problem, prompt neg to set: plackart, belt, breechcloth.
Let yourself be tempted by these 3 tags if you make your shirtless character wear armor, I think you'll appreciate the few models that appear. :)

Several tags to trigger this specific lora:
▣ macrobear
▣ muscular
▣ big muscles
▣ hyper muscules
▣ musclegut
▣ obese
▣ overweigth
▣ bara / barazoku
▣ macro
▣ macro, giga,
▣ macro, terra macro,
▣ planet in hand
▣ crush planet

If you want a very specific body type, I advise you to put the body types you don't want in prompt neg.
It's perfectly possible to make the races you want, just set the weight of the lora to between 0.5 and 0.8 and add bear in neg prompt.

Positive prompt :
cosmic background, cattle, overweight, macro, sitting, space, sitting on planet, horn, plackart, tauren, (3 fingers, hooves:1.3), braided mane, tail tuft, nebula,
<lora:macrobearv3.1:0.5>

Negative prompt :
deformityv6, sitting on ground, boring_e621_fluffyrock_v4, veiny muscles, vein, bear,
Et voilà. :)

If you have any suggestions for making this lora concept even better, please don't hesitate to share them. :)

SDXL version :
the lora responds very well to prompts, thanks to SDXL.
Yes, the lora is almost 900 MB, but that's due to the method I used for training. It gives better detail and increased response to prompts.

~~~~~~~~~~~~~~~~~~~~~~~o(*￣︶￣*)o~~~~~~~~~~~~~~~~~~~~~~~

示例提示词和参数：
Prompt: simple background, canyon, macrobear, front view, bear, fur markings, blue eyes, glowing eyes, broad shoulders, dark grey fur, white markings, teeth, thick neck, big trapezius, armor, armored boots, gauntlets, punch, punch ground, earthquake, bent legs, leaning forward, debris, smoke, crouching, crush, fist crush, impact
<lora:macro_bear_V3_final-000037:1>, (by lindong, by grisser, by markwulfgar, by dramamine)

Negative prompt: deformityv6, watermark, (bad-artist:1.0), (loli:1.2), (worst quality, low quality:1.4), lowres, bad anatomy, bad hands, ((text)), error, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality, normal quality, ((username)), blurry, (extra limbs), bad-artist-anime, (three hands:1.5), (three legs:1.1), (more than two hands:1.4), (more than two legs, :1.2), bad_prompt_version2-neg, boring_e621_fluffyrock_v4, deformityv6, FastNegativeV2

Steps: 20, CFG scale: 7, Sampler: DPM++ 3M SDE Karras, Seed: 1191091475, Size: 1440x1440, Model: indigoFurryMix_v95Realistic, Version: v1.7.0, Pad conds: True, Model hash: b4dbe624d3, "deformityv6: 8455ec9b3d31, FastNegativeV2: a7465e7cc2a2", Denoising strength: 0.45, bad_prompt_version2-neg: 6f35e7dd816a, boring_e621_fluffyrock_v4: f80ed3fee940, "macro_bear_V3_final-000037: 99166df83512", Clip skip: 2