Commit de456f2d authored by lijian6's avatar lijian6
Browse files

Modify README.md


Signed-off-by: lijian6's avatarlijian <lijian6@sugon.com>
parent 5ec77c7d
......@@ -6,11 +6,11 @@
## 模型结构
Vision Transformer先将图像用卷积进行分块以降低计算量,再对每一块进行展平处理变成序列,然后将序列添加位置编码和cls token,再输入多层Transformer结构提取特征,最后将cls tooken取出来通过一个MLP(多层感知机)用于分类。
![img](./doc/vit.png)
![img](./docs/vit.png)
## 算法原理
图像领域借鉴《Transformer is all you need!》算法论文中的Encoder结构提取特征,Transformer的核心思想是利用注意力模块attention提取特征:
![img](./doc/attention.png)
![img](./docs/attention.png)
## 环境配置
### Docker(方法一)
```
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment