扩散模型Diffusion Model
Diffusion Model
解决什么问题What problem it solves
生成图片的主流方法：从一片噪声一步步「擦」出图像。The dominant method for image generation: "wipe" an image out of noise step by step.
这是什么What it is
从随机噪声开始，通过多步去噪逐步生成图像的生成模型。
现在绝大多数图像生成工具都基于它。开源的 Stable Diffusion 系列也是。
A generative model that starts from random noise and produces an image through many denoising steps.
Almost every image-generation tool today, including the open-source Stable Diffusion family, is built on it.
为什么需要知道Why it matters
它是图像生成的技术底座。理解它就理解了为什么生图要等一会儿、以及为什么「步数」是个参数。It is the technical base of image generation. Understanding it explains why generation takes a moment and why "steps" is a parameter.
现在能做什么What you can do now
去导航：找现成的工具、平台、评测与学习资源Directory: ready-made tools, platforms, benchmarks and courses ↗
回到名词地图看第2站的其他 35 个词 →Back to the map: the other 35 terms in station 2 →
出处Sources
- DDPM 去噪扩散概率模型论文Denoising Diffusion Probabilistic Models (DDPM) ↗