PhotoFlow XL

This SDXL ComfyUI workflow has been developed to generate realistic photos with a focus on people.

The following checkpoints are particularly suitable:

SFW

• RealVisXL 5.0: https://civitai.com/models/139562?modelVersionId=789646

• Juggernaut XL: https://civitai.com/models/133005/juggernaut-xl

• HelloWorld XL: https://civitai.com/models/43977/leosams-helloworld-xl

NSFW

• Lustify Endgame: https://civitai.com/models/573152?modelVersionId=1094291

• Big Love XL2: https://civitai.com/models/897413?modelVersionId=1464711

• XXX-Ray v1.1: https://civitai.com/models/1064836?modelVersionId=1305584

Installation

ComfyUI Manager Installation

https://github.com/ltdrdata/ComfyUI-Manager?tab=readme-ov-file#comfyui-manager

Installation of Missing Nodes

Missing nodes in this workflow are highlighted in red. To install them, open the ComfyUI Manager and click on the “Install Missing Custom Nodes” button.

⚠️‼️ComfyUI-KJNodes needed‼️⚠️

It it may happen that the ComfyUI-KJNodes are not included in the list of missing nodes. Simply install the ComfyUI-KJNodes via ComfyUI Manager > Custom Nodes Manager.

After the installation, restart ComfyUI and refresh the browser page.

Recommended Default Values

Sampler: DPM++ SDE GPU (dpmpp_sde_gpu)
Scheduler: normal (⚠️on change, don't forget to change FaceDetailer's own Scheduler, too)
Steps: 25
CFG: 5.6

Recommended Upscalers

• 4xFFHQDAT: adds VERY fine hair and skin details, but smoothes image

• ITF SkinDiffDetail Lite v1: removes smoothness, adds texture, skin details and grain

• UniversalUpscalerV2 Neutral: absolute allround talent, improves texture, especially skin and hair without artificial look or artifacts

Upscaler Downloads

Save in ComfyUI\models\upscale_models directory:

• 4xFFHQDAT: https://openmodeldb.info/models/4x-FFHQDAT

• ITF SkinDiffDetail Lite v1:

  • https://civitai.com/models/1325866/itf-skindiffdetail-lite-v1

  • https://openmodeldb.info/models/1x-ITF-SkinDiffDetail-Lite-v1

• UniversalUpscalerV2 Neutral: https://openmodeldb.info/models/4x-UniversalUpscalerV2-Neutral

💡 If too much detail is lost due to the 4xFFHQDAT upscaler, use UniversalUpscalerV2 Neutral instead.

Ultimate SD Upscale Default Options

Denoise: 0.15 (💡reduce, if the body anatomy is not correct)
Steps: 12
Tiling: no

FaceDetailer

Default Options

Denoise: 0.30 (0.10 to 0.50, depends on image)
Steps: 18 (6 to 25, depends on denoise and image)

💡For close-up shots of the face, the FaceDetailer can be skipped.

SAM Model Installation (if missing)

sam_vit_b_01ec64.pth: ComfyUI Mangager > Model Manager > Search for "ViT-B SAM model"

BBox Detector Installation (if missing)

face_yolov8m.pt: ComfyUI Mangager > Model Manager > Search for "face_yolov8m (bbox)"

VAE Installation

Some checkpoints do not contain a VAE. The official sdxl_vae.safetensors can be downloaded here:

• https://civitai.com/models/296576/sdxl-vae

• https://huggingface.co/stabilityai/sdxl-vae/tree/main

Save it in the directory ComfyUI\models\vae.

TL;DR

The workflow was developed and tested with a NVIDIA RTX 4080 (16 GB VRAM) and 96 GB RAM.

Leave a positive review, if you like the workflow 😊.