Are you sure you want to delete this task? Once this task is deleted, it cannot be recovered.
1207540056@qq.com 5e396003d7 | 2 years ago | |
---|---|---|
configs | 2 years ago | |
images | 2 years ago | |
output/scale_x2/best_model | 2 years ago | |
paddleseg | 2 years ago | |
tools | 2 years ago | |
.gitignore | 2 years ago | |
README.md | 2 years ago | |
compute_classweight.py | 2 years ago | |
predict.py | 2 years ago | |
requirements.txt | 2 years ago | |
train.py | 2 years ago | |
val.py | 2 years ago |
本项目基于paddlepaddle框架复现了ESPNetV2语义分割模型,ESPNetV2利用分组卷积、深度可分离空洞卷积减少模型参数。
论文:
项目参考:
在CityScapes val数据集的测试效果如下表。
steps | opt | image_size | batch_size | dataset | memory | card | mIou | config | |
---|---|---|---|---|---|---|---|---|---|
ESPNetV2 | 120k | adam | 1024x512 | 8 | CityScapes | 32G | 4 | 0.6956 | espnet_cityscapes_1024_512_120k_x2.yml |
硬件: Tesla V100 * 4
框架:
# clone this repo
git clone https://github.com/justld/EspnetV2_paddle.git
cd EspnetV2_paddle
安装第三方库
pip install -r requirements.txt
运行compute_classweight.py文件,注意修改文件内的数据路径,将运行打印的输出结果作为配置文件的损失函数权重。
单卡训练:
python train.py --config configs/espnet_cityscapes_1024_512_120k_x2.yml --do_eval --use_vdl --log_iter 100 --save_interval 1000 --save_dir output
多卡训练:
python -m paddle.distributed.launch train.py --config configs/espnet_cityscapes_1024_512_120k_x2.yml --do_eval --use_vdl --log_iter 100 --save_interval 1000 --save_dir output
output目录下包含已经训练好的模型参数以及对应的日志文件。
python val.py --config configs/espnet_cityscapes_1024_512_120k_x2.yml --model_path output/scale_x2/best_model/model.pdparams
代码结构
├─configs
├─images
├─output
├─paddleseg
│ export.py
│ predict.py
│ README.md
│ README_CN.md
│ requirements.txt
│ setup.py
│ train.py
│ val.py
说明
1、本项目在Aistudio平台,使用Tesla V100 * 4 脚本任务训练120K miou达到69.56%。
2、本项目基于PaddleSeg开发。
相关信息:
信息 | 描述 |
---|---|
作者 | 郎督、胡慧明 |
日期 | 2021年9月 |
框架版本 | Paddle develop |
应用场景 | 语义分割 |
硬件支持 | GPU、CPU |
在线体验 | notebook, Script |
Dear OpenI User
Thank you for your continuous support to the Openl Qizhi Community AI Collaboration Platform. In order to protect your usage rights and ensure network security, we updated the Openl Qizhi Community AI Collaboration Platform Usage Agreement in January 2024. The updated agreement specifies that users are prohibited from using intranet penetration tools. After you click "Agree and continue", you can continue to use our services. Thank you for your cooperation and understanding.
For more agreement content, please refer to the《Openl Qizhi Community AI Collaboration Platform Usage Agreement》