nova-anime-xl-ilv-19.0

Original Model At: https://civitai.com/models/376130/nova-anime-xl?modelVersionId=2940478

Quick Start

Here's some starter code and dependencies to try out with Diffusers library :)<br> Make sure you have correct PyTorch binary compile installed for your GPU.

pip install diffusers==0.39.0 transformers==4.57.6 accelerate==1.14.0 safetensors==0.8.0 pillow==12.2.0

import torch
from diffusers import StableDiffusionXLPipeline

device = "cuda"
pipe = StableDiffusionXLPipeline.from_pretrained(
    "wssdaas/nova-anime-xl-ilv-19.0",
    torch_dtype=torch.bfloat16,
    use_safetensors=True,
).to(device)

prompt = "1girl, shigure ui, safe, anime style, best quality, 4k, absurdres, detailed background, blurry foreground, bokeh, depth of field, volumetric lighting"
negative_prompt = "bad quality, worst quality, bad anaomy, nudity, bad fingers"
steps = 25
cfg = 3.5
width = 768
height = 1344

image = pipe(
    prompt,
    negative_prompt=negative_prompt,
    num_inference_steps=steps,
    guidance_scale=cfg,
    width=width,
    height=height,
).images[0]

image.save("output.png")