Loading
Have a ComfyUI workflow? Get a readable Workflow Doc in your browser.Make a workflow readable.Try it now ↗
This model will help you improve faces in your SDXL image generations.
Model source
Source description
This model will help you improve faces in your SDXL image generations.
Model Type: Textual Inversion (Embedding)
Compatible With: all SDXL models (Pony, Illustrious, ect)
Version: 1.0
File: detfac.safetensors
Trigger Words: detfac
This embedding is a positives-only synthetic textual inversion embedding created through runtime token merging, designed to enhance facial details in SDXL generations. It focuses on photorealistic improvements like sharp eyes, lifelike skin texture with pores, symmetrical features, vibrant expressions, and emotional depth. It counters SDXL's common face smoothness by amplifying realism and intricacy, making it ideal for portraits or scenes with human subjects.
Sources
1 sourceModel artifacts
1 artifactSource excerpts
1 excerptSource context: 332 downloads · Type TextualInversion · Base model SDXL 1.0
In Main Positive Prompt: Simply add "detfac" to your prompt for global facial enhancement, e.g.: "photorealistic portrait of a woman, detfac, masterpiece, 8k". Use weights like (detfac:1.1) to adjust intensity.
With ADetailer Face Models: For ultimate convenience, place "detfac" in ADetailer's positive prompt box during auto-inpainting (e.g., with face_ yolov8s.pt model). It streamlines face fixes by providing targeted detail boosts without cluttering the main prompt—perfect for batch generations or complex scenes where ADetailer detects and refines faces automatically.
Base Prompt Without: Often results in smoothed, less expressive faces.
With detfac: Adds pores, iris details, and symmetry for more lifelike results.
