AryamaR/sdxl_dreambooth_lora_yarn_art_v3

iamno-one/sdxl_dreambooth_lora_yarn_art_v3 is finetuned a SDXL base with Dreambooth Lora for style transfer.

Github repo

Check Git Repo

Model Card for Model ID

from diffusers import DiffusionPipeline, AutoencoderKL
import torch

vae = AutoencoderKL.from_pretrained("madebyollin/sdxl-vae-fp16-fix", torch_dtype=torch.float16)

pipe = DiffusionPipeline.from_pretrained(
    "stabilityai/stable-diffusion-xl-base-1.0",
     vae=vae,
    torch_dtype=torch.float16,
    variant="fp16",
    use_safetensors=True
).to(device)


## adding Lora weight
pipe.load_lora_weights("iamno-one/sdxl_dreambooth_lora_yarn_art_v3")

Sample Prompt

 prompt = "a mother playing with a toddler girl in the park,yarn art style"
negative_prompt = "blurry, low quality, distorted, extra limbs, bad anatomy, deformed face, realistic face"
image = pipe(prompt,height=512, width=512).images[0]
image.show() 

 prompt = "a girl playing with a balloon, yarn art style"
image = pipe(prompt,height=512, width=512).images[0]
image.show()

prompt = "a toddler girl went to zoo,sunny background with lots of trees.The girl is looking at a monkey. Monkey is inside the cage. Toddler girl is standing outside the cage, in yarn art style"
negative_prompt = "blurry, low quality, distorted, extra limbs, bad anatomy, deformed face, realistic face"
image = pipe(prompt,height=512, width=512).images[0]
image.show()