Linaqruf/anime-nouveau-xl-lora

Anime Nouveau XL LoRA

Overview

Anime Nouveau XL LoRA is an innovative LoRA (Low-Rank Adaptation) adapter, meticulously crafted to work seamlessly with Animagine XL 2.0. This model is specifically designed to infuse anime-style images with the richness and intricate ornamentation characteristic of the Anime Nouveau art style. It's perfect for users looking to create anime images that are not just visually stunning but also rich in detailed artistry.

Model Details

• Developed by: Linaqruf
• Model type: LoRA adapter for Stable Diffusion XL
• Model Description: Anime Nouveau XL LoRA is a sophisticated model adapter that enhances the capabilities of Animagine XL 2.0 by adding an Anime Nouveau touch to the images. This style is known for its rich details and ornamental flourishes, offering a unique blend of modern anime art with a classic, detailed aesthetic.
• License: CreativeML Open RAIL++-M License
• Finetuned from model: Animagine XL 2.0

---

🧨 Diffusers Installation

Ensure the installation of the latest diffusers library, along with other essential packages:

pip install diffusers --upgrade
pip install transformers accelerate safetensors

The following Python script demonstrates how to utilize the Anime Nouveau XL LoRA with Animagine XL 2.0. The default scheduler is EulerAncestralDiscreteScheduler, but it can be explicitly defined for clarity.

import torch
from diffusers import (
    StableDiffusionXLPipeline, 
    EulerAncestralDiscreteScheduler,
    AutoencoderKL
)

# Initialize LoRA model and weights
lora_model_id = "Linaqruf/anime-nouveau-xl-lora"
lora_filename = "anime-nouveau-xl.safetensors"

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

# Configure the pipeline
pipe = StableDiffusionXLPipeline.from_pretrained(
    "Linaqruf/animagine-xl-2.0", 
    vae=vae,
    torch_dtype=torch.float16, 
    use_safetensors=True, 
    variant="fp16"
)
pipe.scheduler = EulerAncestralDiscreteScheduler.from_config(pipe.scheduler.config)
pipe.to('cuda')

# Load and fuse LoRA weights
pipe.load_lora_weights(lora_model_id, weight_name=lora_filename)
pipe.fuse_lora(lora_scale=0.6)

# Define prompts and generate image
prompt = "face focus, cute, masterpiece, best quality, 1girl, sketch, monochrome, greyscale, green hair, sweater, looking at viewer, upper body, beanie, outdoors, night, turtleneck"
negative_prompt = "lowres, bad anatomy, bad hands, text, error, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality, normal quality, jpeg artifacts, signature, watermark, username, blurry"

image = pipe(
    prompt, 
    negative_prompt=negative_prompt, 
    width=1024,
    height=1024,
    guidance_scale=12,
    num_inference_steps=50
).images[0]

# Unfuse LoRA before saving the image
pipe.unfuse_lora()
image.save("anime_girl.png")