Pytorch中的卷积、空洞卷积和组卷积 | 小王同学

https://a171232886.github.io/2021/11/21/Pytorch%E4%B8%AD%E7%9A%84%E5%8D%B7%E7%A7%AF%E3%80%81%E7%A9%BA%E6%B4%9E%E5%8D%B7%E7%A7%AF%E5%92%8C%E7%BB%84%E5%8D%B7%E7%A7%AF/

标准卷积Conv2d
最基础的卷积。下面图虽然丑了点，但足够说明问题了。



注： （1） kernel的值在初始化中是随机生成的，可以每个值之间都不一样。
（2）每个通道只对应一个bias值。

The issue links to a blog post about PyTorch convolutions. Review the post content and the repository's structure (likely a Hexo static site). The task is to incorporate this explanation into the site's documentation or blog section. Check the source files for posts or docs, verify the formatting, and ensure the embedded images or diagrams are properly linked.

Scritto dal modello di indicizzazione a partire dal testo della issue.