Realistic Illustration style
LORA Based on my model - Ri-mix - PONY XL
Note: be careful with Neg prompts
Increase LORA strength to get wanted results
My settings for ForgeUI/Automatic1111:
Strength: 0.8-1.0
Sampling method: Euler a
Hires. fix
-
512x768 to 1024x1536
-
None for the upscaler
-
5 for Denoising strength
-
20-40 Sampling steps
-
5-7 CFG Scale
ADetailer - Inpainting
-
ADetailer detector - face_yolov8s.pt
-
Inpaint denoising strength - 2
-
Use separate width/height -1024x1536
POS
zPDXL3, PnyCmicXLPOS, detailxl,
NEG
negativeXL_D, zPDXL-neg, PDXLneg-NoAniPny_-neg, unaestheticXL_bp5,
(Embeddings used in the last image)
For Civitai on-site gen. search:
-
Pony Diffusion V6 XL - V6 (start with this one)
-
Pony PDXL Negative Embeddings - High Quality V3
-
Comic Style PDXL Embeddings - PnyCmicXLPOS
-
Negative Embeds for Pony XL - PDXLneg-NoAniPny
描述:
Recommends strength
训练词语: A illustration in the rimixb style
名称: Rimix_Bravo-f.safetensors
大小 (KB): 56379
类型: Model
Pickle 扫描结果: Success
Pickle 扫描信息: No Pickle imports
病毒扫描结果: Success