import torch
from diffusers import DiffusionPipeline
# switch to "mps" for apple devices
pipe = DiffusionPipeline.from_pretrained("fill-in-base-model", dtype=torch.bfloat16, device_map="cuda")
pipe.load_lora_weights("Erog0291/reton-v3")
prompt = "Astronaut in a jungle, cold color palette, muted colors, detailed, 8k"
image = pipe(prompt).images[0]Reton v3 - Luxury Design LoRA for SD 3.5 Medium
Trigger word: retonstyle
LoRA (rank 32, 1500 steps) fine-tuned on Stable Diffusion 3.5 Medium over a 480-image luxury design dataset (dark-and-gold flyers, website UIs, posters, studio portraits). v3 uses SD3.5's T5-XXL text encoder for much better spelling and layout than the SDXL versions.
import torch
from diffusers import StableDiffusion3Pipeline
pipe = StableDiffusion3Pipeline.from_pretrained(
"adamo1139/stable-diffusion-3.5-medium-ungated", torch_dtype=torch.float16
)
pipe.enable_model_cpu_offload()
pipe.load_lora_weights("Erog0291/reton-v3", weight_name="reton-v3-lora.safetensors")
img = pipe(prompt="retonstyle, luxury event flyer for a Lagos tech summit, dark and gold, elegant serif typography",
num_inference_steps=40, guidance_scale=4.5, height=1024, width=1024).images[0]
Recommended: 40 steps, guidance 4.5, 1024x1024. See samples/ for examples generated with this LoRA.
- Downloads last month
- 20