Kitten Tower

Kitten Tower은 귀여운 스타일에 초점을 맞춘 NoobAI XL 모델입니다.

VAE는 내장되어 있습니다.

v-예측을 지원하는 웹UI를 사용해야 합니다.

• ComfyUI

• reForge

• AUTOMATIC1111 (설정 파일을 체크포인트와 동일한 폴더에 배치하세요)

라이선스

이 모델은 NoobAI-XL 라이선스(개정된 Fair AI 공개 라이선스 1.0-SD) 를 준수하며, 비상업적 용도로만 사용할 수 있습니다. 이 라이선스는 모델, 파생 모델, 또는 모델 생성 제품의 수익화 또는 상업적 사용을 포함하여任何形式의 상업화를 금지합니다.

권장 설정

• 스텝: 25-30

• CFG 스케일: 5-7

• 샘플러: Euler a, Euler

  v-예측 모델을 사용할 경우 "Euler" 사용을 권장합니다. "Euler a"를 사용하면 빨간색, 파란색 등 색상이 강조된 피부나 객체가 생성될 수 있습니다.

• vPred용 스케줄러: Normal, Simple (다른 스케줄러는 이미지 손상 발생)

• 필요 시 ADetailer 사용

긍정적 프롬프트

masterpiece, best quality

부정적 프롬프트

worst quality, bad quality, low quality, lowres, scan artifacts, jpeg artifacts, sketch, light particles

병합 레시피 (Epsilon-pred v1.0)

1. CyberRealistic XL v5와 stable-diffusion-xl-base-1.0의 차이점을 "Add Difference"로 NoobAI-XL Epsilon-pred v1.1에 병합 (체크포인트 A)

2. 체크포인트 A의 CLIP을 NoobAI-XL Epsilon-pred v1.1로 대체 (체크포인트 B)

3. Cat Carrier v3.0를 "Weighted Sum"으로 체크포인트 B에 병합, 블록 (0,0,0,0,0,0,0,0,0.3,0.3,0.3,0,0.3,0.3,0.3,0,0,0.3,0,0) (체크포인트 C)

4. Cat Tower Epsilon-pred v1.2를 "Add Cosine B"로 체크포인트 C에 병합, 블록 (0,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0,0,0,0,0) (체크포인트 D)

5. ioliPonyMix v3.1와 ponyDiffusionV6XL의 차이점을 "Train Difference"로 체크포인트 D에 병합, 블록 (0,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0,0,0,0,0) (체크포인트 E)

6. IllusionBreed v3.0와 NoobAI-XL-Epsilon-pred 1.1-Version의 차이점을 "Train Difference"로 체크포인트 E에 병합, 블록 (0,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.0,0,0,0,0) (체크포인트 F)

7. CatPod v1.0(병합했으나 공개하지 않음)과 ponyDiffusionV6XL의 차이점을 "Train Difference"로 체크포인트 F에 병합, 블록 (0,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0,0,0,0,0) (체크포인트 G)

8. 스타일 LoRA를 체크포인트 G에 병합 (Kitten Tower Epsilon-pred v1.0)

병합 레시피 (vPred v2.0)

1. CyberRealistic XL v5와 stable-diffusion-xl-base-1.0의 차이점을 "Add Difference"로 NoobAI-V-Pred-1.0-Version에 병합 (체크포인트 A)

2. 체크포인트 A의 CLIP을 NoobAI-V-Pred-1.0-Version로 대체 (체크포인트 B)

3. Cat Carrier v3.0를 "Weighted Sum"으로 체크포인트 B에 병합, 블록 (0,0,0,0,0,0,0,0,0.4,0.4,0.4,0,0.4,0.4,0.4,0,0,0.4,0,0) (체크포인트 C)

4. Kitten Tower vpred v1.0를 "Add Cosine B"로 체크포인트 C에 병합, 블록 (0,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0,0,0,0,0) (체크포인트 D)

5. ioliPonyMix v3.1와 ponyDiffusionV6XL의 차이점을 "Train Difference"로 체크포인트 D에 병합, 블록 (0,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0,0,0,0,0) (체크포인트 E)

6. IllusionBreed v3.0와 NoobAI-XL-Epsilon-pred 1.1-Version의 차이점을 "Train Difference"로 체크포인트 E에 병합, 블록 (0,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.0,0,0,0,0) (체크포인트 F)

7. CatPod v1.0(병합했으나 공개하지 않음)과 ponyDiffusionV6XL의 차이점을 "Train Difference"로 체크포인트 F에 병합, 블록 (0,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0,0,0,0,0) (체크포인트 G)

8. 스타일 LoRA를 체크포인트 G에 병합 (체크포인트 H)

9. noob_v_pencil-XL 병합 레시피를 사용해 체크포인트 H에 v_pred 및 ztsnr 키 추가 (Kitten Tower vPred v1.0)

병합 레시피 (vPred v1.0)

1. Cat Tower v1.5 vPred를 "Weighted Sum"으로 NoobAI-V-Pred-1.0-Version에 병합, 블록 (0,0.75,0.75,0.75,0.75,0.75,0.75,0.75,0.75,0.75,0.75,0.75,0.75,0.75,0.75,0,0,0,0,0) (체크포인트 A)

2. ioliPonyMix v3.1와 ponyDiffusionV6XL의 차이점을 "Train Difference"로 체크포인트 A에 병합, 블록 (0,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0,0,0,0,0) (체크포인트 B)

3. IllusionBreed v3.0와 NoobAI-XL-Epsilon-pred 1.1-Version의 차이점을 "Train Difference"로 체크포인트 B에 병합, 블록 (0,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.0,0,0,0,0) (체크포인트 C)

4. CatPod v1.0(병합했으나 공개하지 않음)과 ponyDiffusionV6XL의 차이점을 "Train Difference"로 체크포인트 C에 병합, 블록 (0,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0.2,0,0,0,0,0) (체크포인트 D)

5. 스타일 LoRA를 체크포인트 D에 병합 (체크포인트 E)

6. noob_v_pencil-XL 병합 레시피를 사용해 체크포인트 E에 v_pred 및 ztsnr 키 추가 (Kitten Tower vPred v1.0)