Overview

This guide explains how to generate images similar to mine using WAI-NSFW Illustrious SDXL with AUTOMATIC1111 (1.10.1) and Forge (neo), including my LoRA setup, token merging, and multi-pass ADetailer configuration.

AUTOMATIC1111 (1.10.1)

⚙️ Core Settings

Steps: 40

A higher step count improves refinement and line consistency, especially for tall portrait resolutions.

Sampler: Euler a

Fast, stable, and well-suited for anime and illustration-focused checkpoints.

Schedule Type: Automatic

Automatically selects the optimal scheduling behavior for the sampler.

CFG Scale: 5

Keeps a good balance between prompt adherence and natural variation without oversharpening.

Resolution: 1024x1536

Portrait-oriented resolution optimized for character illustrations and full-body shots.

🧩 Model

Model/Checkpoint: WAI-NSFW-illustrious-SDXL (v16)

👤 ADetailer – Face

Detector Model: face_yolov8n.pt

Confidence: 0.3

Detects faces reliably without false positives.

Dilate / Erode: 4

Mask Blur: 4

Denoising Strength: 0.4

Refines facial features while preserving the original expression and identity.

Inpaint Only Masked: true

Inpaint Padding: 32

ADetailer Version: 25.3.0

Purpose: Subtle facial cleanup and detail enhancement without affecting lighting or composition.

FORGE (NEO)

⚙️ Core Settings

Steps: 20

Lower step count works well with Forge optimizations and token merging.

Sampler: Euler a

Schedule Type: Automatic

CFG Scale: 6

Slightly higher CFG improves prompt accuracy in Forge without excessive stiffness.

Resolution: 800x998

Efficient base resolution that upscales cleanly with post-processing.

Clip Skip: 2

Enhances subject clarity and stylized outlines. If you experience instability in Forge, try setting this to 1.

ENSD: 31337

Ensures consistent output behavior across runs.

Token Merging Ratio: 0.5

Improves performance by merging less-important tokens. If detail loss occurs, reduce to 0.25 or disable it.

🧩 Model & Precision

Model/Checkpoint: WAI-NSFW-illustrious-SDXL (v16)

Diffusion in Low Bits: Automatic (fp16 LoRa)

Module 1: sdxl_vae

The VAE is already embedded, but explicitly setting sdxl_vae may fix some ADetailer-related artifacts.

👤 ADetailer – Face (2nd Pass)

Detector Model: face_yolov8n.pt

Confidence: 0.3

Dilate / Erode: 4

Mask Blur: 4

Denoising Strength: 0.4

Inpaint Only Masked: True

Inpaint Padding: 32

Purpose: Facial refinement after the base pass, maintaining correct identity and proportions.

✋ ADetailer – Hands (3rd Pass)

Detector Model: hand_yolov8n.pt

Prompt: hands, hand, female hand

Confidence: 0.63

Higher confidence reduces incorrect detections.

Dilate / Erode: 4

Mask Blur: 4

Denoising Strength: 0.4

Inpaint Only Masked: True

Inpaint Padding: 32

Use Separate Steps: True

Steps: 13

ADetailer Version: 25.3.0

Purpose: Corrects hand anatomy and finger structure without affecting the rest of the image.

⚠️ Important Note (Very Common Mistake)

Never apply ADetailer on the 1st pass. Doing so may unintentionally alter identity-defining features such as:

• Eye color

• Facial expression

• Face shape

Always run ADetailer after the base generation (and after HiRes, if enabled) to preserve character consistency.

>_ Prompt Recommendations

This is my suggested structure for building a prompt:

1. Quality Tags

   • Focus on tags like masterpiece, best quality, and amazing quality to set the desired level of detail and artistic quality.

2. Character Type & Number

   • Specify whether you want a single character (e.g., 1girl, 1boy) or multiple characters. If there are multiple characters, you can also indicate the number of characters (e.g., 2girls).

3. Trigger Word

   • This is used to activate the lora.

4. Character Details

   • Include specific attributes like hair color, eye color, personality, or specific features (e.g., blonde hair, red eyes, long hair, medium breasts, etc.).

5. Clothing & Outfit

   • Describe the character’s clothing in detail, including colors, fabrics, and specific items like white shirt, red skirt, sneakers, or armor.

6. Pose & View/Camera Position

   • Specify the character's pose (e.g., standing, sitting, jumping) and camera angle (e.g., front view, side view, overhead view). This is essential for guiding the composition of the image.

7. Background & Illumination Details

   • Define the background style (e.g., white background, simple background, urban setting, nature) and include lighting conditions (e.g., soft lighting, backlighting, sunset). This helps set the mood and atmosphere of the scene.

🚫 Negative Prompt Recommendations

NO NSFW Negative Prompt

bad quality, worst quality, worst detail, sketch:1.2, ((watermark, text, logo, signature, artist name, name, labels, artist signaure)), guro, futa, futanari, femboy, nude, naked, nsfw, suggestive, lewd, wardrobe malfunction, sheer clothing, see-through, skirt lift, exposed, exposed breasts, exposed nipples, exposed pussy, see-through clothing, transparent clothing, cameltoe, nipples, nipple bulge, testicles, penis, pussy, colored outline:1.2, bad hands:1.2, bad anatomy:1.2, malformed body:1.2, fused body parts:1.2, multiple navels:1.2, double navels:1.2, extra navel:1.2, ((baby face:1.2, youthful:1.2, young:1.2))

NSFW Negative Prompt

bad quality, worst quality, worst detail, sketch:1.2, ((watermark, text, logo, signature, artist name, name, labels, artist signaure)), censor, mosaic censoring, bad hands:1.2, bad anatomy:1.2, malformed body:1.2, fused body parts:1.2, futa, futanari, multiple navels:1.2, double navels:1.2, extra navel:1.2, ((baby face:1.2, youthful:1.2, young:1.2))

📌 Additional Notes & Tips

This is not a final or exhaustive guide, please feel free to experiment and share suggestions to help improve it. Thank you.