深度学习

CNN and Transformer: What I Should Understand First

整理卷积归纳偏置、注意力机制、视觉任务中的模型结构差异,以及后续复现路线。

8 min readCNNTransformervision

This is a generated placeholder article page. Later, this route can load Markdown or MDX files by slug while keeping the same metadata fields: title, summary, date, category, tags, reading time, and featured status.

Suggested structure: motivation, core ideas, implementation notes, experiments, limitations, and next steps. 中文内容也可以自然混排,用于记录课程、项目复盘、科研阅读和阶段性思考。

// Future MDX slot
export const metadata = {
  title: "CNN and Transformer: What I Should Understand First",
  category: "深度学习"
};