eramth/realism-sdxl

A realism portrait SDXL model with a memory-efficient SDXL VAE that saves about 3GB of RAM with almost no loss of image quality during VAE decoding.

Recommend arguments

step: 20-30, CFG: 2-4

Usage

from diffusers import StableDiffusionXLPipeline
import torch
pipeline = StableDiffusionXLPipeline.from_pretrained("eramth/realism-sdxl",torch_dtype=torch.float16).to("cuda")
# This allows you to generate higher resolution images without much extra VRAM usage.
pipeline.vae.enable_tiling()
image = pipeline(prompt="a beautiful woman",num_inference_steps=25,guidance_scale=2.5).images[0]
image