Stable Diffusion——尝试如何正确的使用【AnimateDiff】生成动图——4种方法测试

这篇具有很好参考价值的文章主要介绍了Stable Diffusion——尝试如何正确的使用【AnimateDiff】生成动图——4种方法测试。希望对大家有所帮助。如果存在错误或未考虑完全的地方,请大家不吝赐教,您也可以点击"举报违法"按钮提交疑问。

测试方案1、传统生成图片方案

给的提示词与正常生成图片的操作一样。

animatediff 提示词 过程,腾讯云 HAI域探秘,stable diffusion

 封面图片生成信息:种子(654742741)

Celebrity face, jet black long hair, beautiful woman, wearing a white shirt, front facing photo of upper body, photo taken, ultra clear, cute, loli, natural black pupils, divine eyes, Chinese style, symmetrical, straight facial features, no stretching, first love, light blue background, tie, campus, table and chair, school uniform, long hair and waist length, smile, dimples, selfies taken from an angle, classmates playing, identification of Class 1 in Grade 2
Negative prompt: Avoid Blemishes,Minimize Wrinkles,ng_deepnegative_v1_75t, (badhandv4:1.5), (worst quality:2),(lower eyelashes:2), (low quality:2), (normal quality:2),(holding:2),(gloves:2),(blue gloves:2),(black gloves:2),lowres, bad anatomy, bad hands, ((monochrome)), ((grayscale)) ,watermark
Steps: 20, Sampler: DPM++ 2M Karras, CFG scale: 7, Seed: 654742741, Size: 512x512, Model hash: 18ed2b6c48, Model: xxmix9realistic_v40, AnimateDiff: "enable: True, model: mm_sd_v15_v2.ckpt, video_length: 16, fps: 8, loop_number: 0, closed_loop: R-P, batch_size: 16, stride: 1, overlap: 4, interp: Off, interp_x: 10, mm_hash: 69ed0f5f", ADetailer model: mediapipe_face_full, ADetailer confidence: 0.3, ADetailer dilate erode: 4, ADetailer mask blur: 4, ADetailer denoising strength: 0.4, ADetailer inpaint only masked: True, ADetailer inpaint padding: 32, ADetailer version: 23.11.1, Version: v1.6.0-2-g4afaaf8a

测试方案2、用语言描绘动作

图片还是断开的,是否我用的词语太多了。再次继续尝试。

animatediff 提示词 过程,腾讯云 HAI域探秘,stable diffusion

(a girl 1.3), (drinking a beverage 1.1), (beverage bottle placed on the table), (celebrity face, jet black long hair, beautiful woman, wearing a white shirt, front of upper body photo, photo taken, ultra clear, cute, loli, natural black pupils, lively eyes, national style, symmetrical, straight facial features, no stretch, first love, light blue background, tie, campus, tables and chairs, school uniform, long hair and waist, huge chest, smile, dimples, elevation selfie 1.2)
Negative prompt: Avoid Blemishes,Minimize Wrinkles,ng_deepnegative_v1_75t, (badhandv4:1.5), (worst quality:2),(lower eyelashes:2), (low quality:2), (normal quality:2),(holding:2),(gloves:2),(blue gloves:2),(black gloves:2),lowres, bad anatomy, bad hands, ((monochrome)), ((grayscale)) ,watermark
Steps: 20, Sampler: DPM++ 2M Karras, CFG scale: 7, Seed: 3742571062, Size: 512x512, Model hash: 18ed2b6c48, Model: xxmix9realistic_v40, AnimateDiff: "enable: True, model: mm_sd_v15_v2.ckpt, video_length: 16, fps: 8, loop_number: 0, closed_loop: R-P, batch_size: 16, stride: 1, overlap: 4, interp: Off, interp_x: 10, mm_hash: 69ed0f5f", ADetailer model: mediapipe_face_full, ADetailer prompt: "(a girl 1.3), (drinking a beverage 1.1), (beverage bottle placed on the table), (celebrity face, jet black long hair, beautiful woman, wearing a white shirt, front of upper body photo, photo taken, ultra clear, cute, loli, natural black pupils, lively eyes, national style, symmetrical, straight facial features, no stretch, first love, light blue background, tie, campus, tables and chairs, school uniform, long hair and waist, huge chest, smile, dimples, elevation selfie 1.2)", ADetailer negative prompt: "Avoid Blemishes,Minimize Wrinkles,ng_deepnegative_v1_75t, (badhandv4:1.5), (worst quality:2),(lower eyelashes:2), (low quality:2), (normal quality:2),(holding:2),(gloves:2),(blue gloves:2),(black gloves:2),lowres, bad anatomy, bad hands, ((monochrome)), ((grayscale)) ,watermark", ADetailer confidence: 0.3, ADetailer dilate erode: 4, ADetailer mask blur: 4, ADetailer denoising strength: 0.4, ADetailer inpaint only masked: True, ADetailer inpaint padding: 32, ADetailer version: 23.11.1, Version: v1.6.0-2-g4afaaf8a

测试方案3、简化用词只描述动作

这回简单的描绘了一下场景以及内容,相对来说好一些。

animatediff 提示词 过程,腾讯云 HAI域探秘,stable diffusion

The little sister who is drinking coffee in the office is sweet, high-definition, delicate, wearing JK, with long black hair, lively eyes, black pupils, pink lips, and the feeling of first love
Negative prompt: Avoid Blemishes,Minimize Wrinkles,ng_deepnegative_v1_75t, (badhandv4:1.5), (worst quality:2),(lower eyelashes:2), (low quality:2), (normal quality:2),(holding:2),(gloves:2),(blue gloves:2),(black gloves:2),lowres, bad anatomy, bad hands, ((monochrome)), ((grayscale)) ,watermark
Steps: 20, Sampler: DPM++ 2M Karras, CFG scale: 7, Seed: 1916071052, Size: 512x512, Model hash: 18ed2b6c48, Model: xxmix9realistic_v40, AnimateDiff: "enable: True, model: mm_sd_v15_v2.ckpt, video_length: 16, fps: 8, loop_number: 0, closed_loop: R-P, batch_size: 16, stride: 1, overlap: 4, interp: Off, interp_x: 10, mm_hash: 69ed0f5f", ADetailer model: mediapipe_face_full, ADetailer prompt: Coffee drinking little sister, ADetailer negative prompt: "Avoid Blemishes,Minimize Wrinkles,ng_deepnegative_v1_75t, (badhandv4:1.5), (worst quality:2),(lower eyelashes:2), (low quality:2), (normal quality:2),(holding:2),(gloves:2),(blue gloves:2),(black gloves:2),lowres, bad anatomy, bad hands, ((monochrome)), ((grayscale)) ,watermark", ADetailer confidence: 0.3, ADetailer dilate erode: 4, ADetailer mask blur: 4, ADetailer denoising strength: 0.4, ADetailer inpaint only masked: True, ADetailer inpaint padding: 32, ADetailer version: 23.11.1, Version: v1.6.0-2-g4afaaf8a

测试方案4、换成动画描述

(((best quality, masterpiece, good anatomy))), eula_in_genshin, cowboy shot, 1girl, blue short hair, black hairband,a leotard with a white top and black bottom, white long sleeves, blue necktie, jewelry on the shoulder, upper body,outdoors, castle, streets, green plants, flowers
Negative prompt: ng_deepnegative_v1_75t, (badhandv4:1.5), (worst quality:2),(lower eyelashes:2), (low quality:2), (normal quality:2),(holding:2),(gloves:2),(blue gloves:2),(black gloves:2),lowres, bad anatomy, bad hands, ((monochrome)), ((grayscale)) ,watermark
Steps: 20, Sampler: DPM++ 2M Karras, CFG scale: 7, Seed: 1542048958, Size: 512x512, Model hash: 18ed2b6c48, Model: xxmix9realistic_v40, AnimateDiff: "enable: True, model: mm_sd_v15_v2.ckpt, video_length: 16, fps: 8, loop_number: 0, closed_loop: R-P, batch_size: 16, stride: 1, overlap: 4, interp: Off, interp_x: 10, mm_hash: 69ed0f5f", ADetailer model: mediapipe_face_full, ADetailer prompt: "(((best quality, masterpiece, good anatomy))), eula_in_genshin, cowboy shot, 1girl, blue short hair, black hairband,a leotard with a white top and black bottom, white long sleeves, blue necktie, jewelry on the shoulder, upper body,outdoors, castle, streets, green plants, flowers", ADetailer negative prompt: "ng_deepnegative_v1_75t, (badhandv4:1.5), (worst quality:2),(lower eyelashes:2), (low quality:2), (normal quality:2),(holding:2),(gloves:2),(blue gloves:2),(black gloves:2),lowres, bad anatomy, bad hands, ((monochrome)), ((grayscale)) ,watermark", ADetailer confidence: 0.3, ADetailer dilate erode: 4, ADetailer mask blur: 4, ADetailer denoising strength: 0.4, ADetailer inpaint only masked: True, ADetailer inpaint padding: 32, ADetailer version: 23.11.1, Version: v1.6.0-2-g4afaaf8a 

 

animatediff 提示词 过程,腾讯云 HAI域探秘,stable diffusion

masterpiece, concept art, dynamic angle, a girl with pink hair blowing a bubble, by WLOP, featured on cgsociety, photorealism, lolita style, soft shadows on the face, sculpted out of candy, [ bubbles, naver fanpop, blonde anime girl with long hair, ( ( ( anime ) ) ), a human-like juicy peach
Negative prompt: ng_deepnegative_v1_75t, (badhandv4:1.5), (worst quality:2),(lower eyelashes:2), (low quality:2), (normal quality:2),(holding:2),(gloves:2),(blue gloves:2),(black gloves:2),lowres, bad anatomy, bad hands, ((monochrome)), ((grayscale)) ,watermark
Steps: 20, Sampler: DPM++ 2M Karras, CFG scale: 7, Seed: 1151319479, Size: 512x512, Model hash: 18ed2b6c48, Model: xxmix9realistic_v40, AnimateDiff: "enable: True, model: mm_sd_v15_v2.ckpt, video_length: 16, fps: 8, loop_number: 0, closed_loop: R-P, batch_size: 16, stride: 1, overlap: 4, interp: Off, interp_x: 10, mm_hash: 69ed0f5f", ADetailer model: mediapipe_face_full, ADetailer prompt: "masterpiece, concept art, dynamic angle, a girl with pink hair blowing a bubble, by WLOP, featured on cgsociety, photorealism, lolita style, soft shadows on the face, sculpted out of candy, [ bubbles, naver fanpop, blonde anime girl with long hair, ( ( ( anime ) ) ), a human-like juicy peach", ADetailer negative prompt: "ng_deepnegative_v1_75t, (badhandv4:1.5), (worst quality:2),(lower eyelashes:2), (low quality:2), (normal quality:2),(holding:2),(gloves:2),(blue gloves:2),(black gloves:2),lowres, bad anatomy, bad hands, ((monochrome)), ((grayscale)) ,watermark", ADetailer confidence: 0.3, ADetailer dilate erode: 4, ADetailer mask blur: 4, ADetailer denoising strength: 0.4, ADetailer inpaint only masked: True, ADetailer inpaint padding: 32, ADetailer version: 23.11.1, Version: v1.6.0-2-g4afaaf8a文章来源地址https://www.toymoban.com/news/detail-792217.html

到了这里,关于Stable Diffusion——尝试如何正确的使用【AnimateDiff】生成动图——4种方法测试的文章就介绍完了。如果您还想了解更多内容,请在右上角搜索TOY模板网以前的文章或继续浏览下面的相关文章,希望大家以后多多支持TOY模板网!

本文来自互联网用户投稿,该文观点仅代表作者本人,不代表本站立场。本站仅提供信息存储空间服务,不拥有所有权,不承担相关法律责任。如若转载,请注明出处: 如若内容造成侵权/违法违规/事实不符,请点击违法举报进行投诉反馈,一经查实,立即删除!

领支付宝红包赞助服务器费用

相关文章

  • Stable Diffusion AnimateDiff-最火文本生成视频插件V3发布

    Stable Diffusion AnimateDiff-最火文本生成视频插件V3发布

    AnimateDiff采用控制模块来影响Stable Diffusion模型,通过大量短视频剪辑的训练,它能够调整图像生成过程,生成一系列与训练视频剪辑相似的图像。简言之,AnimateDiff通过训练大量短视频来优化图像之间的过渡,确保视频帧的流畅性。 与传统的SD模型训练方式不同,AnimateDiff通过

    2024年04月26日
    浏览(12)
  • Stable Diffusion + AnimateDiff运用

    Stable Diffusion + AnimateDiff运用

    1.安装AnimateDiff,重启webui 2.下载对应的模型,最好到c站下载,google colab的资源有可能会出现下载问题 https://civitai.com/models/108836 3.下载完成后,你可以随便抽卡了。 抽卡完成后固定seed,然后打开这个插件,然后通过填写一下参数,点生成就可以开始生成。可以首先固定生成好

    2024年02月08日
    浏览(10)
  • 如何使用Stable Diffusion生成艺术二维码?

    如何使用Stable Diffusion生成艺术二维码?

    物理内存:至少16G(8G直接安装阶段就卡死) N卡:此处我使用GTX 1660 6G (2019年双12购买) windows 11 网络要通畅 git: https://git-scm.com/download/win Python: https://www.python.org/ftp/python/3.10.6/python-3.10.6-amd64.exe CUDA驱动:https://developer.nvidia.com/cuda-toolkit-archive 如何查看下载哪个版本的CUDA驱动?

    2024年02月08日
    浏览(13)
  • 如何使用 ChatGPT 生成 Stable diffusion 提示词

    如何使用 ChatGPT 生成 Stable diffusion 提示词

    http://chat.xutongbao.top StableDiffusion是一款利用深度学习的文生图模型,支持通过使用提示词来产生新的图像,描述要包含或省略的元素。 我在这里引入StableDiffusion算法中的Prompt概念,又被称为提示符。 下面的prompt是用来指导AI绘画模型创作图像的。它们包含了图像的各种细节,

    2024年02月03日
    浏览(9)
  • Stable Diffusion 动画SD-Animatediff V2

    Stable Diffusion 动画SD-Animatediff V2

    AI不仅可以生成令人惊叹的图片,还能给这些图片注入生命,让它们动起来。 这就是AnimateDiff要做的事情,一个神奇的工具,能将静态的AI生成图像转换成动画。 本次介绍基于SD如何实现这个神奇的方法。 进入SD页面在扩展下输入网址 https://github.com/continue-revolution/sd-webui-anim

    2024年02月07日
    浏览(12)
  • Stable Diffusion 动画animatediff-cli-prompt-travel

    Stable Diffusion 动画animatediff-cli-prompt-travel

    基于 sd-webui-animatediff 生成动画或者动态图的基础功能, animatediff-cli-prompt-travel 突破了部分限制,能让视频生成的时间更长,并且能加入controlnet和提示词信息控制每个片段,并不像之前 sd-webui-animatediff 的一套控制全部画面。 动图太大传不上来,凑合看每一帧的图片吧

    2024年02月08日
    浏览(9)
  • 尝试着在Stable Diffusion里边使用SadTalker进行数字人制作

    尝试着在Stable Diffusion里边使用SadTalker进行数字人制作

    首先需要标明的是,我这里是图片说话类型,而且是看了知识星球AI破局俱乐部大航海数字人手册进行操作的。写下这篇文章是防止我以后遗忘。 我使用的基础软件是Stable Diffusion,SadTalker是作为插件放进来的,需要注意的是这对自己的电脑GPU要求比较高,至少需要8G,至少我

    2024年04月09日
    浏览(11)
  • 【AIGC】如何在使用stable-diffusion-webui生成图片时看到完整请求参数

    【AIGC】如何在使用stable-diffusion-webui生成图片时看到完整请求参数

    通过代码调用Stable Diffusion的txt2img、img2img接口时,很多时候都不知道应该怎么传参,比如如何指定模型、如何开启并使用Controlnet、如何开启面部修复等等,在sd-webui上F12看到的请求也不是正式调用SD的请求,所以当引入新插件或需要使用新功能时,怎么传参成了一个大问题,

    2024年02月08日
    浏览(14)
  • 如何让Stable Diffusion正确画手(1)-通过embedding模型优化图片质量

    如何让Stable Diffusion正确画手(1)-通过embedding模型优化图片质量

    都说AI画手画不好手, 看这些是我用stable diffusion生成的图片,小姐姐都很漂亮,但手都千奇百怪,破坏了图片的美感。 其实只需要一个提示词,就能生成正确的手部,看这是我重新生成的效果,每一个小姐姐都有了正确的手部。 首先我们打开LibLibAI模型站或者Civitai模型站,

    2024年02月16日
    浏览(13)
  • stable diffusion模型初尝试

    stable diffusion模型初尝试

    提示:文章写完后,目录可以自动生成,如何生成可参考右边的帮助文档 文章目录 前言 一、 目录 文章目录 前言 一、stable diffusion - webUI的下载安装以及配置 二、模型训练 刚开始尝试基于stable diffusion跑模型,试着做一些小项目,网上能查到的资料着实有限,部分甚至需要科

    2024年02月12日
    浏览(9)

觉得文章有用就打赏一下文章作者

支付宝扫一扫打赏

博客赞助

微信扫一扫打赏

请作者喝杯咖啡吧~博客赞助

支付宝扫一扫领取红包,优惠每天领

二维码1

领取红包

二维码2

领红包