Text-to-Image
Diffusers
lora
sdxl
streetwear
t-shirt-design
graphic-design
ukiyo-e
puff-print
typography
Instructions to use adarrshDev/streetwear-decal-sdxl-lora with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Diffusers
How to use adarrshDev/streetwear-decal-sdxl-lora with Diffusers:
pip install -U diffusers transformers accelerate
import torch from diffusers import DiffusionPipeline # switch to "mps" for apple devices pipe = DiffusionPipeline.from_pretrained("stabilityai/stable-diffusion-xl-base-1.0", dtype=torch.bfloat16, device_map="cuda") pipe.load_lora_weights("adarrshDev/streetwear-decal-sdxl-lora") prompt = "TOK_DECAL, TOK_ARCHIVAL_DECAL, Japanese woodblock dual cranes with botanical pine waves and mineral green ink, ukiyo-e aesthetic, clean vector contours, screenprint plastisol ready, isolated on black background, zero mockup, zero human" image = pipe(prompt).images[0] - Inference
- Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- Draw Things
- DiffusionBee
import torch
from diffusers import DiffusionPipeline
# switch to "mps" for apple devices
pipe = DiffusionPipeline.from_pretrained("stabilityai/stable-diffusion-xl-base-1.0", dtype=torch.bfloat16, device_map="cuda")
pipe.load_lora_weights("adarrshDev/streetwear-decal-sdxl-lora")
prompt = "TOK_DECAL, TOK_ARCHIVAL_DECAL, Japanese woodblock dual cranes with botanical pine waves and mineral green ink, ukiyo-e aesthetic, clean vector contours, screenprint plastisol ready, isolated on black background, zero mockup, zero human"
image = pipe(prompt).images[0]π¨ Streetwear Decal SDXL LoRA (Golden Universe Edition)
Fine-tuned Stable Diffusion XL (SDXL 1.0) LoRA trained on the Golden 450-Decal Streetwear Universe. Specifically engineered to generate isolated, screenprint-ready 2D graphic t-shirt decals on solid black backgrounds with zero human models, zero faces, zero collar seams, and zero mockups.
π Verified Quality Audit & Zero-Human Guarantee
Following an automated OpenCV YuNet neural computer-vision audit () and chromatic skin-tone analysis across all training data:
| Metric | Result |
|---|---|
| Detected Faces in Dataset | 0 / 450 (100% Passed) |
| Human Models / Anatomy | 0 / 450 (100% Purged) |
| Garment Seams / Collar Tags | 0 / 450 (Stripped) |
| Training Steps | 1,500 Steps (Tesla T4 GPU) |
| Precision | FP16 Mixed Precision, 8-bit AdamW |
| LoRA Rank | 32 (Rank-32 weights, 185.96 MB) |
π Trigger Words & Aesthetics
Always invoke the core trigger token along with specific sub-aesthetic tokens:
- : Japanese Ukiyo-e woodblock, Edo-period mineral ink, classical botanical/wildlife motifs (Kamisaka Sekka, Hokusai, De Stijl).
- : 90s Acid Surf 3D high-relief puff print, textured foam lettering.
- : Vintage racing, speedway rally typography, collegiate varsity athletics.
- : Cyber-sigilism, liquid chrome metallic tribal contours, dark techwear vectors.
- : Monochrome Japanese manga ink, dark anime street calligraphy, sharp screenprint linework.
- : Scandinavian minimalist fine line art, delicate botanical silhouettes.
- : Outdoor adventure and mountain topography line art, Grand Prix circuit line maps.
- : Street luxury crests, baroque chain link typography.
π» How to Use in Diffusers
- Downloads last month
- 29
Model tree for adarrshDev/streetwear-decal-sdxl-lora
Base model
stabilityai/stable-diffusion-xl-base-1.0