动漫图像扩散生成模型 EimisAnimeDiffusion

本模型使用具有高质量和丰富细节的动漫图像训练而来。

Sample generations

本模型擅长生成动漫人物和风景。
一些示例如下：

Positive:a girl, Phoenix girl, fluffy hair, war, a hell on earth, Beautiful and detailed explosion, Cold machine, Fire in eyes, burning, Metal texture, Exquisite cloth, Metal carving, volume, best quality, normal hands, Metal details, Metal scratch, Metal defects, masterpiece, best quality, best quality, illustration, highres, masterpiece, contour deepening, illustration,(beautiful detailed girl),beautiful detailed glow
Negative:lowres, bad anatomy, ((bad hands)), text, error, ((missing fingers)), cropped, jpeg artifacts, worst quality, low quality, signature, watermark, blurry, deformed, extra ears, deformed, disfigured, mutation, censored, ((multiple_girls))
Steps: 20, Sampler: DPM++ 2S a, CFG scale: 8, Seed: 4186044705/4186044707, Size: 704x896

Positive:(1girl), cute, walking in the park, (night), full moon, north star, blue shirt, red skirt, detailed shirt, jewelry, autumn, dark blue hair, shirt hair, (magic:1.5), beautiful blue eyes
Negative: lowres, bad anatomy, ((bad hands)), text, error, ((missing fingers)), cropped, jpeg artifacts, worst quality, low quality, signature, watermark, blurry, deformed, extra ears, deformed, disfigured, mutation, censored, ((multiple_girls))
Steps: 35, Sampler: Euler a, CFG scale: 9, Seed: 296195494, Size: 768x960

Positive:night , ((1 girl)), alone, masterpiece, 8k wallpaper, highres, absurdres, high quality background, short hair, black hair, multicolor hair, beautiful frozen village, (full bright moon), blue dress, detailed dress, jewelry dress, (magic:1.2), blue fire, blue eyes, glowing eyes, fire, ice goddess, (blue detailed beautiful crown), electricity, blue electricity, blue light particles
Negative: lowres, bad anatomy, ((bad hands)), text, error, ((missing fingers)), cropped, jpeg artifacts, worst quality, low quality, signature, watermark, blurry, deformed, extra ears, deformed, disfigured, mutation, censored, ((multiple_girls))
Steps: 20, Sampler: DPM++ 2S a Karras, CFG scale: 9, Seed: 2118767319, Size: 768x832

想要生成惊艳的背景图片吗？没问题：

Positive: above clouds, mountains, (night), full moon, castle, huge forest, forest between mountains, beautiful, masterpiece
Negative: lowres, bad anatomy, ((bad hands)), text, error, ((missing fingers)), cropped, jpeg artifacts, worst quality, low quality, signature, watermark, blurry, deformed, extra ears, deformed, disfigured, mutation, censored, ((multiple_girls))
Steps: 20, Sampler: DPM++ 2S a Karras, CFG scale: 9, Seed: 83644543, Size: 896x640

使用方式

from modelscope.utils.constant import Tasks
from modelscope.pipelines import pipeline
import cv2

pipe = pipeline(task=Tasks.text_to_image_synthesis,
                model='dienstag/EimisAnimeDiffusion_1.0v',
                model_revision='v1.0.0')

prompt = "a girl, Phoenix girl, fluffy hair, war, a hell on earth, Beautiful and detailed explosion, Cold machine, Fire in eyes, burning, Metal texture, Exquisite cloth, Metal carving, volume, best quality, normal hands, Metal details, Metal scratch, Metal defects, masterpiece, best quality, best quality, illustration, highres, masterpiece, contour deepening, illustration,(beautiful detailed girl),beautiful detailed glow"
neg_prompt = "lowres, bad anatomy, ((bad hands)), text, error, ((missing fingers)), cropped, jpeg artifacts, worst quality, low quality, signature, watermark, blurry, deformed, extra ears, deformed, disfigured, mutation, censored, ((multiple_girls))"
output = pipe({'text': prompt, 'negative_prompt': neg_prompt})
cv2.imwrite('result.png', output['output_imgs'][0])

License

这个模型是开放的，所有人都可以使用，CreativeML OpenRAIL-M许可证进一步规定了权利和使用。CreativeML OpenRAIL许可证规定了：

你不能使用这个模型来故意产生或分享非法或有害的输出或内容。
作者对你产生的输出结果没有任何权利，你可以自由使用它们，并对它们的使用负责，但不得违反许可证中的规定。
你可以重新发布模型权重，并在商业上作为一项服务使用该模型。如果你这样做，请注意你必须包括与许可证中相同的使用限制，并将CreativeML OpenRAIL-M的副本分享给你的所有用户（请完全仔细阅读许可证）。请在这里阅读完整的许可证。