Stable Diffusion nsfw Prompts

Today we are using the following Stable Diffusion checkpoints

realisticVisionV60B1

link to civtai or on huggingface.co

For version 6.0 it is recommended to use with VAE (to improve generation quality and get rid of artifacts): https://huggingface.co/stabilityai/sd-vae-ft-mse-original

positive:

RAW photo, subject, 8k uhd, dslr, soft lighting, high quality, film grain, Fujifilm XT3

negative:

(deformed iris, deformed pupils, semi-realistic, cgi, 3d, render, sketch, cartoon, drawing, anime), text, cropped, out of frame, worst quality, low quality, jpeg artifacts, ugly, duplicate, morbid, mutilated, extra fingers, mutated hands, poorly drawn hands, poorly drawn face, mutation, deformed, blurry, dehydrated, bad anatomy, bad proportions, extra limbs, cloned face, disfigured, gross proportions, malformed limbs, missing arms, missing legs, extra arms, extra legs, fused fingers, too many fingers, long neck

config:

Euler A or DPM++ SDE Karras (25+ steps) / DPM++ 2M SDE (50+ steps)

CFG Scale 3,5 – 7

Hires. fix with 4x-UltraSharp upscaler

Denoising strength 0.25-0.45

Upscale by 1.1-2.0

Clip Skip 1-2

ENSD 31337

Resolutions (use lower resolution if you get a lot of mutations and stuff like that)

• Face Portrait: 896×896
• Portrait: 896×896, 768×1024
• Half Body: 768×1024, 640×1152
• Full Body: 896×896, 768×1024, 640×1152, 1024×768, 1152×640

Recommended Workflow

Images can be generated with or without Hires.Fix, but it will help improve the generation quality significantly. In some cases it is strictly recommended to use Hires.Fix, namely when generating full body and half body images (note: you can also use Restore Faces or ADetailer).

Recommended Hires.Fix Parameters
Sampler: DPM++ SDE Karras or DPM++ 2M SDE
Denoising steps: 10+ (DPM++ SDE Karras) / 20+ (DPM++ 2M SDE (notice. the lower the value of hires steps at a given sampler, the stronger the skin texture and the higher the chance of getting artifacts))
Denoising strength: 0.1-0.3
Upscaler: 4x-UltraSharp / 4x_NMKD-Superscale-SP_178000_G or another
Upscale by: 1.1-2.0+