StableDiffusionBook
中文社区共建的 Stable Diffusion 实战百科:从 SDWebUi 安装到 LoRA 训练
如果你读中文、想动手跑图而不是只看概念,这本 wiki 值得先过一遍:安装、报错、ControlNet、五类训练都有现成路径,省掉大量翻英文 issue 的时间。分角色推荐页也很实用,二次元用户直接被指向 NovelAI,设计师则拿到 @秋葉aaaki 的视频教程与一键启动器。软肋同样清楚——部分章节已归档、可能落后于上游工具版本,社区链接里不少群组和创作者资源需要你自己验证是否还活着。把它当起点读,别当成随版本更新的官方手册。
Verified 7d ago · liveness 63/100 · cite: rightaichoice.com/tools/stablediffusionbook
- 中文用户从零安装并跑通 Stable Diffusion WebUi
- 插画师与设计师把 AI 出图接进实际项目工作流
- 准备训练 LoRA、DreamBooth 等自定义风格模型的高级用户
- 想搞清 NovelAI 泄漏事件与社区模型谱系的人
- 需要官方技术支持或 SLA 的团队
- 要找 API、企业级部署或商业授权方案的买家
- 追求与上游工具版本同步的最新参数文档的用户
We scan live Reddit threads, YouTube comments, X posts, G2 reviews and other communities — and hand you an honest verdict in under a minute.
- Honest verdict, not marketing
- Real pros & cons from real users
- Attributed quotes with receipts
3 free scans · no card needed
Skip StableDiffusionBook if you need version-synced parameters for upstream tools, or you want a vendor that hosts models and compute for you — this is a community-maintained wiki, and some chapters are flagged as archived.
文档本身免费且采用 GFDL 协议,但真正跑图所需的本地 GPU 硬件成本由你自己承担
文档免费开放,成本几乎为零,适合预算敏感的个人学习者和中小工作室;代价是学习路径要自己走完,且部分章节已归档,需要额外花时间对照上游工具自行校准。若你需要付费换取时效性与支持,应将预算放到商业教程或厂商服务上。
In short
StableDiffusionBook — 中文社区共建的 Stable Diffusion 实战百科:从 SDWebUi 安装到 LoRA 训练. Best for 中文用户从零安装并跑通 Stable Diffusion WebUi, 插画师与设计师把 AI 出图接进实际项目工作流, 准备训练 LoRA、DreamBooth 等自定义风格模型的高级用户. Free to use.
What people actually say about StableDiffusionBook — is it worth it?
We ran a structured research pass across product reviews, community discussions, and post-purchase forum threads to surface the patterns vendors won't publish themselves. Below: the recurring strengths, the hidden costs people mention most, and the cohort that consistently regrets adopting this tool.
7 mentions across 2 sources (YouTube, GitHub), 1 more we could not attribute · researched Sep 29, 2026.
Weighted by the 8 posts each of 2 sources contributed.
- +Task-based structure (start, install, debug, draw, train, reference) matches how users actually work
- +Documents the NovelAI lineage and both leaked datasets (53.66 GB, 124.54 GB) clearly
- +Covers five training pipelines: LoRA, Textual Inversion, Hypernetwork, DreamBooth, AestheticGradients
- +Maintainers close reported issues fast — WebUi sync closed same day in July 2024
- +Free under GFDL with open contribution, no paywall or account required
- −ComfyUI issue open since April 2023 with no resolution, despite ecosystem billing
- −English translation still pending since March 2023, so bilingual claim is aspirational
- −Archived sections may lag upstream SDWebUi, risking broken install commands
- −Roughly four open issues and a maintainer/translator call signal volunteer fragility
- −No hosted models or compute — full setup still needs your own GPU and storage
- • Your own GPU and electricity for training and inference
- • 53–124 GB downloads for leaked-model packs referenced in the lineage chapter
- • Time cost of volunteer-maintained docs going stale on archived pages
Viability Score
How well maintained and how widely used is StableDiffusionBook? Built from what the vendor actually publishes (docs, changelog, tutorials, integrations, pricing), whether the site is live, and how much real users discuss it. How we calculate this
Last calculated: September 2026
How we score →Key Features
- SDWebUi 安装与运行的分步指南
- GPU 相关错误与常见运行报错排查
- Txt2Img 与 Img2Img 工作流讲解
- 提示工程与提示词书写指南
- ControlNet 实战教程
- LoRA 训练完整流程
- Textual Inversion 训练指南
- Hypernetwork 训练指南
- DreamBooth 训练指南
- AestheticGradients 训练指南
- Stable Diffusion 与 NovelAI 谱系及泄漏事件说明(53.66 GB / 124.54 GB 两个数据包)
- 按用户类型分角色推荐资源(设计师 / 二次元 / 老手)
- 模型源与社区论坛链接汇总(Hugging Face、Civitai、Reddit、NGA、Discord)
- 中英双语文档(English / 简体中文)
- GFDL 协议下的开放共建文档
About StableDiffusionBook
StableDiffusionBook 是一份开放共建的中文(含英文)实战文档,教你把 Stable Diffusion 真正用进工作流。它围绕 Stable Diffusion WebUi(SDWebUi)展开:从安装与运行、GPU 相关错误排查,到提示工程、Txt2Img / Img2Img、ControlNet 实战,再到 LoRA、Textual Inversion、Hypernetwork、DreamBooth、AestheticGradients 五类训练流程,并把 ComfyUI 等相关工具作为生态的一部分介绍。文档按任务而非功能组织,分为起步于此、开始之前、安装配置、调试程序、绘画指南、训练模型、参考资料七块。它少见地把 Stable Diffusion 与 NovelAI 的谱系讲清楚了——包括 NAI 泄漏事件的两个数据包(53.66 GB 与 124.54 GB),以及这些泄漏模型如何成为后来大多数社区模型的训练起点。它还按用户类型分角色推荐资源:设计师/美工推荐 @秋葉aaaki 的视频教程、一键启动器与蓝图节点编辑器,二次元用户则直接指向 NovelAI。适合想动手跑图的中文用户,尤其是插画师、设计师和准备训练自定义风格模型的人。注意:部分章节标注为归档内容,可能落后于上游工具版本;文档本身不托管模型、也不提供算力。
Behind the Verdict
StableDiffusionBook 的价值不在“最新”,而在“成体系 + 中文 + 免费”。它把散落在 GitHub discussions、Reddit、NGA 和一堆 B 站教程里的知识,按任务重排成一条路径:起步于此讲术语和新闻,开始之前讲入门要求,安装配置给 WebUi 的安装与运行步骤,调试程序专门收 GPU 相关和常见报错,绘画指南从概论、提示工程走到 Txt2Img / Img2Img / ControlNet,训练模型依次覆盖 LoRA、Textual Inversion、Hypernetwork、DreamBooth、AestheticGradients,参考资料则把模型源和社区入口集中成一页速查。 最强的一段是 NovelAI 谱系:它明确写出 NAI 泄漏的两个数据包体积(Part 1 为 53.66 GB、含 7GB 和 4GB 两种模型;Part 2 为 124.54 GB),并指出现在大部分社区模型都基于泄漏模型训练而来,因此画风油腻、风格趋同、不服指令,而 NovelAI 后来用 SDXL 做的第三代模型则风格多变、更听话。这段对理解今天的风格同质化很有用,别处不容易一次读到。 分角色推荐资源页是另一个务实设计:设计师/美工直接拿到 @秋葉aaaki 的视频教程、一键启动器和蓝图节点编辑器,二次元用户被指向 NovelAI,老手则继续读归档文件。 缺点是诚实的:文档自述“许多技术文档虽然有些可能已过时,但还是提供了有价值的技术背景”,也就是说参数和版本细节不能当权威依据;它也不托管模型、不提供算力;主地址在图片或外链失效时建议换网络,Vercel 版本在大陆不可用。它适合当学习路线和排错起点,不适合当与上游同步的参考手册。
Researching StableDiffusionBook? Get your full AI stack in 60 seconds.
Free, no signup — tell us your goal and get tools matched to your budget & existing stack.
Real-world workflow fit
Concrete scenarios for the personas StableDiffusionBook actually fits — and what changes day-one when you adopt it.
按“安装配置”里的 WebUi 安装与运行步骤在本机跑通环境,遇到报错时跳到“调试程序”查 GPU 相关和常见错误,再用“绘画指南”的提示工程与 Txt2Img 模式出第一张图。
Outcome: 用一份中文文档走完从安装到出图的完整路径,不必在英文 issue 里逐条排查。
先判断数据集规模与目标,再到“训练模型”按顺序尝试 LoRA 训练,必要时用 Textual Inversion 或 DreamBooth 补充,最后回到绘画指南验证出图效果。
Outcome: 拿到一条从数据集准备到模型训练再到验证的现成路径,减少试错。
阅读“Stable Diffusion 和 NovelAI 的关系”一节,对照泄漏事件 Part 1(53.66 GB)与 Part 2(124.54 GB)的说明,理解为何多数社区模型风格趋同。
Outcome: 建立对当前模型谱系和风格同质化成因的基本认知,再决定用哪类模型。
Use Cases
- 在本机安装并跑通 Stable Diffusion WebUi
- 为出图效果调提示词,学会书写与实战技巧
- 在自己的数据集上训练 LoRA 等自定义风格模型
- 用 ControlNet 做受控的图像编辑与构图
- 排查 GPU 显存与运行性能问题
- 搞清 NovelAI 泄漏事件与社区模型的谱系关系
- 按身份(设计师 / 二次元用户)找到合适的教程与社区入口
Models Under the Hood
as of 2026-09-21
Limitations
文档自述部分技术文档“可能已过时”,章节被标注为归档内容,可能落后于上游工具版本,参数细节不适合作为权威依据。它不托管模型、不提供算力或成品素材,纯属信息性文档。访问上,若图片或外链资源打不开需更换网络,源站返回 404/300/302/502 时项目方建议提交 issue 移除失效内容;Vercel 版本在中国大陆不可用。社区链接与创作者资源需自行验证是否仍然活跃。
as of 2026-09-21
Where the pricing makes sense
The company stage and team size where StableDiffusionBook's pricing actually pencils out — and where peers do it cheaper.
文档免费开放,成本几乎为零,适合预算敏感的个人学习者和中小工作室;代价是学习路径要自己走完,且部分章节已归档,需要额外花时间对照上游工具自行校准。若你需要付费换取时效性与支持,应将预算放到商业教程或厂商服务上。
Setup time & first value
How long it actually takes to get something useful out of StableDiffusionBook — broken out by persona, not the marketing-page minute.
纯阅读的话,几分钟就能翻完“起步于此”并定位到需要的章节;如果是照着“安装配置”在自己机器上跑通 SDWebUi,加上环境与模型下载,通常要花上数小时到一整天,具体取决于硬件和网络。
Switching to or from StableDiffusionBook
How to bring data in from common predecessors and how to get it back out — written for the switcher, not the buyer.
- →From 官方英文文档或 GitHub issue:改用本 wiki 的中文安装、调试与训练章节作为起点,再按需回查上游细节
- →From 零散 B 站视频教程:把视频当补充,用本 wiki 的分章节文本梳理完整路径
- →From ComfyUI 工作流:参考文档中作为生态部分介绍的 ComfyUI 内容,与 SDWebUi 路径对照
- ↗To 上游工具官方文档:当需要与当前版本同步的参数与接口细节时,回到 AUTOMATIC1111 stable-diffusion-webui 等官方来源
- ↗To 商业教程或厂商服务:当你需要时效性、技术支持或算力托管时
Integrations
Resources & Guides
Tutorials & Learning
YouTube returned 6 videos for “StableDiffusionBook”, and we withheld 6: 6 could not be judged, because “StableDiffusionBook” is a single word that other videos use for other things. We are showing none, because we could not prove any of them are about StableDiffusionBook.
Official links
Tools that pair well with StableDiffusionBook
Common stack mates teams adopt alongside StableDiffusionBook, with the specific reason each pairing earns its keep.
Featured Head-to-Head Comparisons
Stablediffusionbook vs Surge Ai
Don't treat this as a head-to-head. StableDiffusionBook is a free, Chinese-language field manual for getting Stable Diffusion WebUi running, writing prompts, and training LoRA/DreamBooth/Textual Inversion styles — you pay nothing and get no compute, no API, and no SLA. Surge AI is the opposite purchase entirely: a staffed vendor selling credentialed doctor/lawyer/engineer human feedback, red teaming, and benchmarks like GDP.pdf and ComplexConstraints to labs that need citable numbers. If you're an illustrator or hobbyist trainer, open the wiki today. If you're training a frontier model and need expert post-training data, book a scoping call with Surge.
Stablediffusionbook vs Praktika
These aren't competitors — they don't belong in the same buying decision. StableDiffusionBook is a free, community-maintained Chinese reference for getting Stable Diffusion WebUi running and training LoRA/DreamBooth/Textual Inversion models; it's the right pick if your problem is image-generation workflows and you read Chinese. Praktika is a ~$8/month mobile app for speaking practice with AI tutors like Raika and Min-Jun, aimed at learners who already have vocabulary but freeze out loud. If you need image models, take the free wiki; if you need speaking reps, take Praktika. Note the wiki's archived sections can lag upstream tool versions, and the app's free tier caps speaking.
Stablediffusionbook vs Sakana Ai
These are not competitors and you should never be choosing between them. StableDiffusionBook is a free Chinese-language wiki that teaches you to install Stable Diffusion WebUi, debug GPU errors, run ControlNet and train LoRA/DreamBooth models — its only cost is your time, and its weakness is that some chapters are archived and drift behind upstream tool versions. Sakana AI is a Tokyo lab selling enterprise contracts to Japanese banks, the defense ministry and intelligence buyers who legally cannot move data offshore, with no published price and no self-serve signup. If you have a budget and a problem, only one of these will even answer your email.
Stablediffusionbook vs Adobe Firefly Services
If you are deciding where to put money and engineering hours, the answer is that these are not substitutes. StableDiffusionBook costs nothing and gets a Chinese-speaking designer or hobbyist from zero to a running SDWebUi install and a trained LoRA — reach for it when your problem is 'I don't know how to start.' Adobe Firefly Services is what you buy when your problem is 'we need generative fill/extend at volume inside our CMS, with SOC2/FedRAMP and indemnified training data.' Buying Firefly does not teach you SDWebUi, and reading the book does not give you an enterprise API. Pick the one that matches your actual blocker.
Stablediffusionbook vs Genspark
These two aren't really competitors — StableDiffusionBook is a free Chinese-language wiki for the Stable Diffusion image-generation community, and Genspark is a commercial AI workspace for research, documents, and no-code agents. If you're a Chinese-speaking illustrator trying to install SDWebUi, train a LoRA, or debug GPU errors, StableDiffusionBook is the resource you want (just note some chapters are archived and lag upstream versions). If you need cited research summaries, AI Slides/Docs/Sheets, or no-code agents, StableDiffusionBook won't help you at all — that's Genspark's territory. They can coexist in a stack, but nobody is choosing one over the other.
Stablediffusionbook vs Undermind
These aren't competitors — a buyer would never weigh them in one decision. If you're a Chinese-speaking illustrator trying to install SDWebUi, fix a GPU error, or train a LoRA, StableDiffusionBook answers that for free, with the caveat that archived sections may lag current tool versions. If you're a researcher or R&D team running an exhaustive literature review where citation traceability matters, Undermind is the tool, and the ~2.9-minute search time is the entry fee. Pick the one that matches your problem; there is no trade-off between them.
Stablediffusionbook vs The New Black
These are not competitors and no buyer should be choosing between them. StableDiffusionBook is a free, Chinese-language handbook for people who want to install Stable Diffusion WebUi, fix GPU errors, write prompts, run ControlNet, and train LoRA or DreamBooth models themselves. The New Black is a paid, productized fashion studio where a brand hands over a sketch or product photo and gets back garment renders, on-model imagery, tech packs with grading and a bill of materials, and social-ready video. Pick StableDiffusionBook if you want to learn and own your image-generation pipeline for free; pick The New Black if you run an apparel or accessory label and need a finished, factory-ready deliverable — and note its 2026 shift to a four-tier plan with rollover credits and unlimited team seats on paid tiers.
Stablediffusionbook vs Runway Gen 4
These aren't competitors — one is a free Chinese-language manual for running Stable Diffusion locally, the other is a credit-metered cloud video studio. If you want to learn SDWebUi, ControlNet, and how to train LoRA or DreamBooth models yourself, StableDiffusionBook costs nothing and covers exactly that (with the caveat that some chapters are archived and may lag upstream versions). If you need finished video — text-to-video, image-to-video, frame-propagation edits in Aleph 2.0, or agent-built ad campaigns with regional localization — you're buying Runway Gen-4, and you should budget for credits rather than expect the 125 free ones to carry a real project. Pick by output: stills and custom models you control vs. video assets you ship on a timeline.
Alternatives to StableDiffusionBook
View allThinkdiffusion
Cloud GPU workspace for running Stable Diffusion, ComfyUI, Forge, Fooocus, and Kohya without local hardware.
Tensor.art
Community-driven AI art platform with diffusion models and granular controls.
Artbreeder
Breed and remix images with AI gene sliders for endless visual exploration.
Frequently Asked Questions
Categories
Best-of guides
Used StableDiffusionBook? Help shape our editorial sentiment research.