T
traeai
Sign in

概念

Diffusion Models

别名:扩散模型

一类基于随机微分方程(SDE)的生成模型,通过逐步去噪从噪声分布生成数据。

已跟踪 3 条高相关材料

TraeAI 观察

相关材料

已收录 3 条与 Diffusion Models 相关的内容,按评分排序。

Diffusion Models for Video Generation

Diffusion Models for Video Generation

Lil'Log3690 字 (约 15 分钟)
85

The article provides an in-depth analysis of diffusion models in video generation, exploring their core mechanisms, architecture design, and recent developments.

入选理由:视频生成比图像生成更复杂,需要更强的时间一致性

FeaturedArticle#Diffusion Models#Video Generation中文
Where the Score Lives: What Wavelets Reveal About Diffusion Models

Where the Score Lives: What Wavelets Reveal About Diffusion Models

Microsoft Research9623 字 (约 39 分钟)
78

A wavelet-based interpretable parameterization of the score function reveals that local interactions most strongly impact denoising performance, and wavelets yield greater gains at higher resolutions; this framework serves as a theoretical probe—not a replacement—for trained models.

入选理由:提出三种可解析求解的wavelet-based score模型:独立基线、带通(band-tide)、局部耦合模型

FeaturedVideo#Diffusion Models#Wavelet Analysis#Generative Models#Interpretable AI英文
Continuous Latent Diffusion Language Model

paper: https://t.co/qgUyX7D1Mr

Continuous Latent Diffusion Language Model

AK(@_akhaliq)50 字 (约 1 分钟)
72

CLDLM leverages continuous latent space modeling to significantly improve generation speed and quality, ideal for high-resolution text generation tasks.

入选理由:CLDLM 使用连续潜在变量,使语言生成过程更平滑,减少离散化误差。

FeaturedTweet#Diffusion Models#Language Modeling#Continuous Latent Space#AI Generation#Natural Language Processing英文

跨材料问答 · Diffusion Models

回答基于:Diffusion Models 相关 3 条材料
    0 / 500

    AI may generate inaccurate information. Please verify important content.