PaddleClas/docs/zh_CN/models/PVTV2.md

28 lines
1.0 KiB
Markdown
Raw Blame History

This file contains ambiguous Unicode characters!

This file contains ambiguous Unicode characters that may be confused with others in your current locale. If your use case is intentional and legitimate, you can safely ignore this warning. Use the Escape button to highlight these characters.

# PVTV2
---
## 目录
* [1. 概述](#1)
* [2. 精度、FLOPS 和参数量](#2)
<a name='1'></a>
## 1. 概述
PVTV2 是 VisionTransformer 系列模型,该模型基于 PVTPyramid Vision Transformer改进得到PVT 模型使用 Transformer 结构构建了特征金字塔网络。PVTV2 的主要创新点有1. 带 overlap 的 Patch embeding2. 结合卷积神经网络3. 注意力模块为线性复杂度。[论文地址](https://arxiv.org/pdf/2106.13797.pdf)。
<a name='2'></a>
## 2. 精度、FLOPS 和参数量
| Models | Top1 | Top5 | Reference<br>top1 | Reference<br>top5 | FLOPS<br>(G) | Params<br>(M) |
|:--:|:--:|:--:|:--:|:--:|:--:|:--:|
| PVT_V2_B0 | 0.705 | 0.902 | 0.705 | - | 0.53 | 3.7 |
| PVT_V2_B1 | 0.787 | 0.945 | 0.787 | - | 2.0 | 14.0 |
| PVT_V2_B2 | 0.821 | 0.960 | 0.820 | - | 3.9 | 25.4 |
| PVT_V2_B3 | 0.831 | 0.965 | 0.831 | - | 6.7 | 45.2 |
| PVT_V2_B4 | 0.836 | 0.967 | 0.836 | - | 9.8 | 62.6 |
| PVT_V2_B5 | 0.837 | 0.966 | 0.838 | - | 11.4 | 82.0 |
| PVT_V2_B2_Linear | 0.821 | 0.961 | 0.821 | - | 3.8 | 22.6 |