Artificial Intelligence

引导扩散

Guided Diffusion

A generative modeling technique that uses a diffusion process to create data samples while being guided by a conditional signal (such as a text prompt or a specific image feature) to control the output. / 一种生成建模技术，利用扩散过程创建数据样本，同时通过条件信号（如文本提示或特定图像特征）进行引导以控制输出。

Knowledge type: 工艺技术
First proposed or discovered
Proposer: Turkey Farm member
Creator: 农场主
Last updated: 2026-11-09 20:05:58

Sub-knowledge points

Forward Diffusion ProcessReverse Denoising ProcessClassifier GuidanceText-to-Image Generation

Higher-order knowledge points

Generative Adversarial Networks (GANs)Variational Autoencoders (VAEs)Diffusion Models

Main parameters and performance indicators

Guidance Scale: Controls the strength of the influence from the guidance signal.Noise Schedule: The predefined schedule for adding Gaussian noise to the data.

Open problems

Mode Collapse: The model may generate limited varieties of samples.Computational Cost: The process is computationally intensive.

Related literature

ranked by community stars

A Scalable Pipeline Combining Procedural 3D Graphics and Guided Diffusion for Photorealistic Synthetic Training Data Generation in White Button Mushroom SegmentationOpen source

引导扩散

Sub-knowledge points

Higher-order knowledge points

Main parameters and performance indicators

Open problems

Related products and links