![](https://img-blog.csdnimg.cn/b069d6304c1248e3ab499e52a9df5d22.png?x-oss-process=image/resize,m_fixed,h_224,w_224)
论文笔记
文章平均质量分 90
组会复习捏
bulibuli蛋
nuaa->seu bulibuli蛋
展开
-
Transformer在CV上的应用-论文总结
整理了读过的一些Transformer相关的论文原创 2023-03-22 17:04:17 · 416 阅读 · 0 评论 -
【Swin Transformer 论文笔记】Swin Transformer: Hierarchical Vision Transformer using Shifted Windows
swin transformer 是一种计算机视觉的通用主干框架。原创 2022-10-20 23:52:17 · 598 阅读 · 1 评论 -
【ViT 论文笔记】AN IMAGE IS WORTH 16X16 WORDS: TRANSFORMERS FOR IMAGE RECOGNITION AT SCALE
We show that this reliance on CNNs is not necessary and a pure transformer applied directly to sequences of image patches can perform very well on image classification tasks原创 2022-10-19 16:00:15 · 551 阅读 · 1 评论 -
【DETR 论文笔记】 End-to-End Object Detection with Transformers
Two ingredients are essential for direct set predictions in detection:(1) a set prediction loss that forces unique matching between predicted and ground truth boxes;(2) an architecture that predicts (in a single pass) a set of objects and models their原创 2022-10-19 15:40:37 · 959 阅读 · 1 评论 -
【Transformer 论文笔记】Attention is all you need
但是如果输入的序列太长,则是截取左边的内容,把多余的直接舍弃。以Encoder-Decoder框架为例,输入Source和输出Target内容是不一样的,比如对于英-中机器翻译来说,Source是英文句子,Target是对应的翻译出的中文句子,Attention发生在Target的元素Query和Source中的所有元素之间。设置8个attention表示不同方向上的注意力,得到8个z0,那就将多个版本的x拼接称为一个长向量,然后用一个全连接网络,即乘以一个矩阵,就能得到一个短的x向量。原创 2022-10-19 15:13:29 · 459 阅读 · 1 评论