Animagine XLをGoogle Colabで試す

Animagine XLはSDXL 1.0からの派生モデルです。

モデル

アニメスタイルの画像を厳選して学習されたモデル。

Linaqruf/animagine-xl · Hugging Face

Google Colabで試す

すぐに試す

• 「ランタイム」から「すべてのセルの実行」で実行

• ランタイムはGPUに設定されています。

もし設定されていなければ「ランタイム」▶「ランタイム設定」▶「T4 GPU」▶「保存」

Google Colaboratory

---

コード

ライブラリのインストール▼

!pip install -q --upgrade diffusers invisible_watermark transformers accelerate safetensors

パイプラインでAnimagine XLモデルを準備▼

from diffusers import DiffusionPipeline, EulerAncestralDiscreteScheduler
import torch

# model = "stabilityai/stable-diffusion-xl-base-1.0"
model = "Linaqruf/animagine-xl"

pipe = DiffusionPipeline.from_pretrained(
    model, 
    torch_dtype=torch.float16, 
    use_safetensors=True, 
    variant="fp16")

pipe.scheduler = EulerAncestralDiscreteScheduler.from_config(pipe.scheduler.config)
pipe.to("cuda")

# cpuモード
# pipe.enable_model_cpu_offload()

画像生成▼

import matplotlib.pyplot as plt

prompt = "masterpiece, best quality, girl, cute, gray hair, long, yellow eyes, black dress, upper body"
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"

# デフォルトサイズ(1024x1024)

image = pipe(
    prompt, 
    negative_prompt=negative_prompt, 
    guidance_scale=12,
    num_inference_steps=25
    ).images[0]

image.save("output.png")

plt.imshow(image)
plt.axis('off')

---

生成画像

ネガティブは「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」で統一しています

masterpiece, best quality, girl cute gray hair long yellow eyes black dress upper body

masterpiece, best quality, face focus, boy, cool, magenta hair, blue eyes, shirt, blue sky

masterpiece, best quality, face focus, girl, cute, pink hair, ponytail, deep eyes, school uniform

masterpiece, best quality, face focus, girl, cool, rainbow hair, starry eyes

masterpiece, best quality, face focus, girl, cute, gray hair, red hat, red eyes, red clothes

masterpiece, best quality, face focus, girl, cute, black hair, dark eyes, clary loli, lolita, one-piece cape, embroidery, plaid, detective uniform, dress

masterpiece, best quality, face focus, 3d figure, girl, cute, black hair, dark eyes, clary loli, lolita, one-piece cape, embroidery, plaid, detective uniform, dress