import torch
from diffusers import DiffusionPipeline
# switch to "mps" for apple devices
pipe = DiffusionPipeline.from_pretrained("xinsir/controlnet-union-sdxl-1.0", dtype=torch.bfloat16, device_map="cuda")
pipe.load_lora_weights("jarod2212/SelfieDynamics_XL")
prompt = "-"
image = pipe(prompt).images[0]Selie Dynamics XL LoRA
- Prompt
- -
- Prompt
- -
- Prompt
- -
- Prompt
- -
- Prompt
- -
- Prompt
- -
- Prompt
- -
Model description
Trigger: sdynxl Recommended weight: 0.70 – 0.85 ⚠️ Do NOT exceed 1.0 — above 1.0 it may glitch (arm distortions, broken perspective)
⭐ Description SelfieDynamics XL is an SDXL LoRA designed to create dynamic selfie compositions with:
extended arm
half‑body framing
tilted camera angles
close‑up perspective
natural motion and energy
vertical 9:16 format
This LoRA does not overwrite identity. 👉 It adds only the selfie dynamics, without forcing ethnicity, hairstyle, makeup, or facial structure.
It is trained to be strong, stable, and highly generalizable.
⭐ Training Details Base model: SDXL 1.0
Dataset: 50 curated 9:16 images
Dynamic poses, extended arm, half‑body
Mixed ethnicities, hairstyles, makeup
All subjects fully dressed
Minimal tagging for maximum generalization
Rank 32, Alpha 16
Learning Rate 0.0008
Weight Decay 0.0001
1800 steps
Optimized to enhance selfie dynamics without overriding the base model.
⭐ Best Use Cases Perfect for generating:
realistic selfies
social‑media‑style portraits
vertical compositions
tilted angles
expressive, energetic poses
natural arm extension
PROMPTS EXAMPLE:
sdynxl, dynamic selfie, arm extended, half-body, 9:16 vertical, tilted camera angle, natural light, fully dressed
sdynxl, dynamic selfie, African woman, arm extended, half-body, 9:16 vertical, curly hair in motion, natural makeup, wearing a long-sleeve top, fully dressed, outdoor daylight
sdynxl, dynamic selfie, Latina woman, arm extended, half-body, 9:16 vertical, diagonal angle, playful expression, long straight dark hair, soft glam makeup, wearing a fitted long-sleeve blouse, fully dressed
sdynxl, dynamic selfie, East Asian woman, arm extended, half-body, 9:16 vertical, cute expression, short bob haircut, minimal makeup, wearing a high-neck long-sleeve top, fully dressed
sdynxl, dynamic selfie, Nordic blonde woman, arm extended, half-body, 9:16 vertical, joyful smile, long wavy blonde hair, glossy skin makeup, wearing a turtleneck sweater, fully dressed
sdynxl, dynamic selfie, Irish redhead woman, arm extended, half-body, 9:16 vertical, playful expression, freckles, wavy red hair, natural makeup, wearing a buttoned cardigan, fully dressed
Trigger words
You should use sdynxl to trigger the image generation.
Download model
Download them in the Files & versions tab.
- Downloads last month
- 13
Model tree for jarod2212/SelfieDynamics_XL
Base model
xinsir/controlnet-union-sdxl-1.0