Sign In

[LECO] Monster meat

516

Updated: Mar 1, 2025

objectsdarkfoodmeat

Download

2 variants available

Type

LoRA

Stats

588

109

18

Reviews

Published

Jul 23, 2023

Base Model

SD 1.5

Hash

AutoV2
11BB273EAE

Trigger Words

meat
default creator card background decoration

toretate

This LoRA is a demonic meat.

use for dark fantasy meal sean.

for Illustrious v0.1

This is a demonic foods (re create for Illustrious v0.1)

This is LECO, good feeling weight is 2.0 ~2.5.

trigger word: meat

wieght: 1 ~ 2

LECO Parameters

- target: "meat"
  positive: "meat, monster, purple, tentacles, multi colored, eyes, blue, yellow, black, insect, dripping, glowing"
  action: "enhance"
  guidance_scale: 2.0
  resolution: 768
  dynamic_resolution: true
  batch_size: 3

below: SuperMerger LoRA merge parameters

ratio: 2
Blocs: 1,0,0,0,0,0,0,0,0,0,1,1,1,1,1,1,1,1,1,1

Be careful when setting Strength to 2 or higher, as it tends to produce significantly darker images.

SD1 版のものを Illustrious v0.1 で作り直したもの

LECOのパラメータ

- target: "meat"
  positive: "meat, monster, purple, tentacles, multi colored, eyes, blue, yellow, black, insect, dripping, glowing"
  action: "enhance"
  guidance_scale: 2.0
  resolution: 768
  dynamic_resolution: true
  batch_size: 3

また、SuperMergerで強度などを補正済みです

ratio: 2
Blocs: 1,0,0,0,0,0,0,0,0,0,1,1,1,1,1,1,1,1,1,1

Strength を 2以上にするとかなりダーク寄りの絵が出てしまうので注意?

ダークファンタジーの食事風景にでもご利用ください


for SD 1

This is a demonic foods.

This is LECO, good feeling weight is 2.0 ~2.5.

If you are going to use it with human, use Regional Prompter or Regional tools. ( If not used with the regional tool, it would often fill the entire image)

trigger word: meat

wieght: 2.0 ~ 2.5 (if Regional tool, use 1~2)

LECO parameters

- target: "meat"
  positive: "meat, monster, tentacles, purple, blue, multi colored, insect, sticky"
  action: "enhance"
  guidance_scale: 1.0
  resolution: 768
  dynamic_resolution: false
  batch_size: 2
- train:
  precision: "bfloat16"
  noise_scheduler: "ddim"
  iterations: 200
  lr: 1
  optimizer: "Prodigy"
  lr_scheduler: "constant"
  max_denoising_steps: 20
  network:
  type: "lierla"
  rank: 4
  alpha: 1.0
  training_method: "full"