# SOLOv2

**Repository Path**: SearchSource/SOLOv2

## Basic Information

- **Project Name**: SOLOv2
- **Description**: No description available
- **Primary Language**: Unknown
- **License**: Apache-2.0
- **Default Branch**: master
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 1
- **Forks**: 0
- **Created**: 2020-04-27
- **Last Updated**: 2020-12-26

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# SOLOv2
The code is an unofficial pytorch implementation of [SOLOv2: Dynamic, Faster and Stronger](https://arxiv.org/abs/2003.10152)


## Install
Please check [SOLOv1](https://github.com/WXinlong/SOLO/blob/master/docs/INSTALL.md) for installation instructions.

## Training
Follows the same way as SOLOv1.

single GPU: 
```
python tools/train.py configs/solov2/solov2_r101_3x.py
```
multi GPU (for example 8): 
```
./tools/dist_train.sh configs/solov2/solov2_r101_3x.py 8
```
## Weights
Trained model can be download in [Google drive](https://drive.google.com/file/d/1v8SCK5LAlbE6X4dnpl2DOaWfakwFzENI/view?usp=sharing)
[BaiduYun](https://pan.baidu.com/s/125AB_xW-GA4anphs1IOAvg) 提取码: qw4e
## Results
After training 36 epochs(3x) on the coco dataset using the resnet-101 backbone, the mAP is 39.5 on COCO test-dev2017 dataset. In the original paper, the model achieves 39.7 after 72 epochs(6x).

<img src="AP.jpg">

## Visualization（1 epoch)

<img src="solov2.png" width="2000">