跳到主要内容
Learn Motion
探索课程产品能力价格

下载 Learn Motion

用手机扫描二维码,下载 App。

App Store

扫码从 App Store 下载 Learn Motion打开商店页面

Google Play

扫码从 Google Play 下载 Learn Motion打开商店页面
登录
Learn Motion

How Text-to-Video AI Works Under the Hood

1From Prompt to Conditioning Signal2Latent Diffusion: Generating Video in Compressed Space3Spatio-Temporal Attention: Keeping Frames Consistent4Where the Pipeline Breaks: Failure Modes and Their Causes
Spatio-Temporal Attention: Keeping Frames Consistent

Attention Inside a Single Frame

1 / 5
Take one frame of the latent video and cut it into a grid of patches. Each patch is just a small bundle of numbers, and on its own it has no idea what it is part of.
0:00 / 0:00

Spatial self-attention operates within one frame. The frame's latent is divided into patches, and each patch produces a query, a key, and a value. Every patch compares its query against all other patches' keys, and the resulting weights decide how much information it pulls from each value. This is how a region that only sees a patch of fur can still represent a whole animal: it gathers evidence from the patches around it. Without this exchange, each patch would be denoised on its own local evidence, and the frame would look like unrelated tiles rather than one scene.

参考资料

  1. [1]Attention Is All You Needarxiv.org
上一页1 / 5下一页

Learn Motion

生成课程,认真学会。

运营方:Wuhan Daoyin Technology Co., Ltd.

联系我们: [email protected]
隐私政策用户服务协议

© 2026 Learn Motion