README.md

metadata

library_name: diffusers

Generate Genshin Impact LandScape style image by lora tuned on stable-diffusion-xl

Install

pip install git+https://github.com/huggingface/diffusers.git peft

import torch
from diffusers import (
    StableDiffusionXLPipeline,
    AutoencoderKL,
)

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

model_path = "stabilityai/stable-diffusion-xl-base-1.0"
pipe = StableDiffusionXLPipeline.from_pretrained(
    model_path, torch_dtype=torch.float16, vae=vae
)
pipe.to("cuda")
pipe.load_lora_weights("svjack/Genshin-Impact-LandScape-lora-sd-xl-rk4")

Generate Genshin Mondstadt LandScape Image

prompt = "European, green coniferous tree, yellow coniferous tree, rock, creek, sunny day, pastel tones, 3D"
image = pipe(prompt=prompt,
             num_inference_steps=50, guidance_scale=5.0,
            cross_attention_kwargs={"scale": 0.7}
            ).images[0]
image

Generate Genshin Mondstadt LandScape Clear Image

prompt = "European, green coniferous tree, yellow coniferous tree, rock, creek, sunny day, pastel tones, 3D"
image = pipe(prompt=prompt,
             negative_prompt = "blurred, uneven, messy, foggy",
             num_inference_steps=50, guidance_scale=5.0,
            cross_attention_kwargs={"scale": 0.7}
            ).images[0]
image

Generate Genshin Liyue LandScape Image

prompt = "Chinese, yellow deciduous wood, orange deciduous wood, rock, sunny day, pastel tones, 3D"
image = pipe(prompt=prompt,
             num_inference_steps=50, guidance_scale=5.0,
            cross_attention_kwargs={"scale": 0.7}
            ).images[0]
image

Generate Genshin Liyue LandScape Clear Image

prompt = "Chinese, yellow deciduous wood, orange deciduous wood, rock, sunny day, bright, 3D"
image = pipe(prompt=prompt,
             num_inference_steps=50, guidance_scale=5.0,
            cross_attention_kwargs={"scale": 0.7}
            ).images[0]
image