PaddleClas/docs/en/models/RedNet_en.md

22 lines
1.0 KiB
Markdown
Raw Normal View History

2021-06-10 23:54:00 +08:00
# RedNet series
2021-12-07 10:31:53 +08:00
---
## Catalogue
2021-06-10 23:54:00 +08:00
2021-12-07 10:31:53 +08:00
* [1. Overview](#1)
2021-12-07 10:39:56 +08:00
* [2. Accuracy, FLOPs and Parameters](#2)
2021-06-10 23:54:00 +08:00
2021-12-07 10:31:53 +08:00
<a name='1'></a>
## 1. Overview
2021-06-10 23:54:00 +08:00
2021-12-07 10:31:53 +08:00
In the backbone of ResNet and in all bottleneck positions of backbone, the convolution is replaced by Involution, but all convolutions are reserved for channel mapping and fusion. These carefully redesigned entities combine to form a new efficient backbone network, called Rednet. [paper](https://arxiv.org/abs/2103.06255).
2021-06-10 23:54:00 +08:00
2021-12-07 10:31:53 +08:00
<a name='2'></a>
2021-12-07 10:39:56 +08:00
## 2. Accuracy, FLOPs and Parameters
2021-06-10 23:54:00 +08:00
| Model | Params (M) | FLOPs (G) | Top-1 (%) | Top-5 (%) |
|:---------------------:|:----------:|:---------:|:---------:|:---------:|
| RedNet26 | 9.2 | 1.7 | 75.95 | 93.19 |
| RedNet38 | 12.4 | 2.2 | 77.47 | 93.56 |
| RedNet50 | 15.5 | 2.7 | 78.33 | 94.17 |
| RedNet101 | 25.7 | 4.7 | 78.94 | 94.36 |
| RedNet152 | 34.0 | 6.8 | 79.17 | 94.40 |