상세 정보
추천
v5.0
v4.0α
v3.0
v2.0
v1.0
LibrasIllustriousXL

LibrasIllustriousXL

9.8K
105
3.6K
#애니메이션
#스타일
#merge
#checkpoint xl
#checkpoint merging
#sdxl 1,0
#illustriousXL
#Illustrious

※ 以前のバージョンの説明はAbout this versionから確認してください。

※ Please refer to "About this version" for details on the previous version.

 

noobaiの系列、illustriousの系列を問わず新しいモデルのいくつかを軸に一からマージをし直しました。そのためV4以前とは全く違う挙動をします。

新しいモデルを取り入れたため縦長ならば少し大きいサイズの出力でも問題がありません。(横長は胴が長くなりやすいです)

I have merged several new models, regardless of whether they belong to the Noobai or Illustrious series. As a result, the behavior is completely different from V4 and earlier versions.

Due to the incorporation of new models, vertically long images can now be generated with relative stability, even at slightly larger sizes. (However, in horizontally long images, the torso tends to appear elongated.)

 

目が崩れやすいので余裕があればADetailerの顔への適用をオススメします。

また確認できる問題点として暗闇の描写がやや苦手です。

Since the eyes can easily become distorted, if possible, it is recommended to apply ADetailer to the face.

One noticeable issue is that the model struggles somewhat with depicting darkness.

 

  • 推奨する品質プロンプト/Recommended quality prompt

masterpiece, best quality, 
  • 推奨するネガティブプロンプト/Recommended negative prompt

worst quality, low quality, realistic, mutated hands, light particles, jpeg artifacts, lowres, signature, twitter username, company name, logo, sidelighting, heart-shaped pupils, 

推奨するプロンプトをまず使用し、そこから調整をしてください。

また、特定のサンプラーや特定のスケジュール、CFGスケールが7を超えるとよくないためサンプルで使用した出力設定も参考にしてください。

例としてDPM++ 2M SDE HeunではNormalやUniformのスケジュールタイプを使用すると強烈にノイズが乗ります。

First use the recommended prompt, and then adjust it from there.

Additionally, certain samplers, specific schedules, and CFG scales above 7 do not work well, so please refer to the output settings used in the samples.

For example, with DPM++ 2M SDE Heun, using the Normal or Uniform schedule types results in excessive noise.

 

VAE内蔵済みです。

The VAE is already integrated.

 

  • サンプルで使用した出力設定/Output settings used for the samples

Sampling method:28

Output size:1040*1520

Schedule type:DPM++ 2M SDE Heun

Sampling steps:SGM Uniform

CFG Scale:5 ※7以上は推奨しません/A CFG scale above 7 is not recommended

Random number generator source:GPU

Emphasis mode:no norm

Eta noise seed delta:31337

CLIP skip SDXL:2

 

  • サンプルを出力した環境/Environment used to output samples

reForge

RTX3060

 

  • マージに使用したモデル/Models used for merging

Magic_ILL

waiNSFWIllustrious_v120

hikariNoobVPred_120

illustriousXL10_v10に手を改善するloraをマージしたモデル/A model merged with a LoRA that improves hand depiction, based on IllustriousXL10_v10

noobai0.75をベースに画風の学習をした自作のモデル/A custom model trained on art style, based on Noobai 0.75

songmix_v201

 

マージに使用したモデル及び、学習のベースとなったモデルの制作者達に感謝いたします。

I would like to express my gratitude to the creators of the models used for merging and the models that served as the base for training.

 

  • merge recipe

Magic_ILL x (1-alpha) + waiNSFWIllustrious_v120 x alpha = A
1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1
A x (1-alpha) + hikariNoobVPred_120 x alpha = B
0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0
B x (1-alpha) + illustriousXL1.0の手改善モデル x alpha = C
0,0,0,0.4,0,0.4,0,0.4,0,0,0.4,0,0,0,0,0,0.4,0.4,0,0
C x (1-alpha) + noobai0.75をベースに画風の学習をした自作のモデル x alpha = D
0,0,0.6,0,0,0.6,0,0,0.6,0.6,0,0.6,0.6,0,0,0.6,0,0,0.6,0
D x (1-alpha) + songmix_v201 x alpha = E
0,0,0,0,0,0,0,0,0.2,0.4,0.4,0.6,0.6,0.2,0,0,0,0,0,0
E.fp16 x (1-alpha) + C x alpha = LibrasIllustriousXL_V5
0,0.4,0.4,0.4,0.4,0.4,0.4,0.4,0.4,0.4,0.4,0.4,0.4,0.4,0.4,0.4,0.4,0.4,0.4,0.4
번역문 보기

평점 및 리뷰

-- /5
0 개의 평점

충분한 평가나 댓글을 받지 못했습니다.

O
공고
2024-10-05
모델 게시
2025-03-24
모델 정보 업데이트
모델 상세정보
유형
Checkpoint
게시 날짜
2025-03-24
기본 모델
Illustrious
버전 소개

※ 以前のバージョンの説明はAbout this versionから確認してください。

※ Please refer to "About this version" for details on the previous version.

noobaiの系列、illustriousの系列を問わず新しいモデルのいくつかを軸に一からマージをし直しました。そのためV4以前とは全く違う挙動をします。

新しいモデルを取り入れたため縦長ならば少し大きいサイズの出力でも問題がありません。(横長は胴が長くなりやすいです)

I have merged several new models, regardless of whether they belong to the Noobai or Illustrious series. As a result, the behavior is completely different from V4 and earlier versions.

Due to the incorporation of new models, vertically long images can now be generated with relative stability, even at slightly larger sizes. (However, in horizontally long images, the torso tends to appear elongated.)

目が崩れやすいので余裕があればADetailerの顔への適用をオススメします。

また確認できる問題点として暗闇の描写がやや苦手です。

Since the eyes can easily become distorted, if possible, it is recommended to apply ADetailer to the face.

One noticeable issue is that the model struggles somewhat with depicting darkness.

  • 推奨する品質プロンプト/Recommended quality prompt

masterpiece, best quality, 
  • 推奨するネガティブプロンプト/Recommended negative prompt

worst quality, low quality, realistic, mutated hands, light particles, jpeg artifacts, lowres, signature, twitter username, company name, logo, sidelighting, heart-shaped pupils, 

推奨するプロンプトをまず使用し、そこから調整をしてください。

また、特定のサンプラーや特定のスケジュール、CFGスケールが7を超えるとよくないためサンプルで使用した出力設定も参考にしてください。

例としてDPM++ 2M SDE HeunではNormalやUniformのスケジュールタイプを使用すると強烈にノイズが乗ります。

First use the recommended prompt, and then adjust it from there.

Additionally, certain samplers, specific schedules, and CFG scales above 7 do not work well, so please refer to the output settings used in the samples.

For example, with DPM++ 2M SDE Heun, using the Normal or Uniform schedule types results in excessive noise.

VAE内蔵済みです。

The VAE is already integrated.

  • サンプルで使用した出力設定/Output settings used for the samples

Sampling method:28

Output size:1040*1520

Schedule type:DPM++ 2M SDE Heun

Sampling steps:SGM Uniform

CFG Scale:5 ※7以上は推奨しません/A CFG scale above 7 is not recommended

Random number generator source:GPU

Emphasis mode:no norm

Eta noise seed delta:31337

CLIP skip SDXL:2

  • サンプルを出力した環境/Environment used to output samples

reForge

RTX3060

  • マージに使用したモデル/Models used for merging

Magic_ILL

waiNSFWIllustrious_v120

hikariNoobVPred_120

illustriousXL10_v10に手を改善するloraをマージしたモデル/A model merged with a LoRA that improves hand depiction, based on IllustriousXL10_v10

noobai0.75をベースに画風の学習をした自作のモデル/A custom model trained on art style, based on Noobai 0.75

songmix_v201

マージに使用したモデル及び、学習のベースとなったモデルの制作者達に感謝いたします。

I would like to express my gratitude to the creators of the models used for merging and the models that served as the base for training.

  • merge recipe

Magic_ILL x (1-alpha) + waiNSFWIllustrious_v120 x alpha = A
1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1
A x (1-alpha) + hikariNoobVPred_120 x alpha = B
0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0
B x (1-alpha) + illustriousXL1.0の手改善モデル x alpha = C
0,0,0,0.4,0,0.4,0,0.4,0,0,0.4,0,0,0,0,0,0.4,0.4,0,0
C x (1-alpha) + noobai0.75をベースに画風の学習をした自作のモデル x alpha = D
0,0,0.6,0,0,0.6,0,0,0.6,0.6,0,0.6,0.6,0,0,0.6,0,0,0.6,0
D x (1-alpha) + songmix_v201 x alpha = E
0,0,0,0,0,0,0,0,0.2,0.4,0.4,0.6,0.6,0.2,0,0,0,0,0,0
E.fp16 x (1-alpha) + C x alpha = LibrasIllustriousXL_V5
0,0.4,0.4,0.4,0.4,0.4,0.4,0.4,0.4,0.4,0.4,0.4,0.4,0.4,0.4,0.4,0.4,0.4,0.4,0.4
허가 범위
모델 출처: civitai

1. 재게시된 모델의 권리는 원 제작자에게 있습니다.

2. 모델 원작자가 모델을 인증받으려면 공식 채널을 통해 SeaArt.AI 직원에게 문의하세요. 저희는 모든 창작자의 권리를 보호하기 위해 노력합니다. 인증하러 이동

창작 허가 범위
온라인 생방송
혼합 진행
다운로드 허용
상업적 허가 범위
생성된 이미지를 판매하거나 상업적 목적으로 사용 가능
모델의 재판매 또는 융합 후 판매 허용
QR Code
SeaArt 앱 다운로드
모바일에서 AI 창작 여정을 계속하세요