base model: STABLE DIFFUSION XL v1.0 / PONY by AstraliteHeart
SD checkpoint: dreamshaperXL10_alpha2xl10.safetensor
input: prompt and image
workflow: ComfyUI
creative operator: STSchwanitz
format: 1920 x 1080 px / 1088 x 1088 px
final retouch: yes Photoshop
Update 2025
SDXL + LoRAs
Using nodes „IPAdapter Unified Loader FaceID“ with „FACEID PLUS V2“ / „IPAdapter FaceID“
Note 1:
Importance of strength (strength_model and strength_clip):
strength_model: Defines how strongly LoRA influences image generation. A value of 1.0 is standard, but values between 0.5 and 0.8 are often used.
strength_clip: Defines how strongly the LoRA influences the text encoder. This value usually remains at 1.0, but can also be adjusted if the text prompts do not deliver the desired results.
Note 2:
Check your embeded python version – if you are updated to 3.12 there is already an „insightface.whl“ for 3.12 version just search for:
„insightface-0.7.3-cp312-cp312-win_amd64.whl“
.
.
PONY by AstraliteHeart + LoRAs
No post examples:
.
More to come.
Status of 2024-02-02
SDXL + LoRAs
Shooting with virtual model Antheia
Click to enlarge to 1920 x 1080 px
.
.
.
.
Briefing:
Consistent character with futuristic fashion in cyber city set