search

qubvel / segmentation_models.pytorch

Semantic segmentation models with 500+ pretrained convolutional and transformer-based backbones.
https://smp.readthedocs.io/
MIT License
9.1k stars 1.62k forks source link

the method of adding new networks #882

Open wangluoWSG opened 1 month ago

wangluoWSG commented 1 month ago

How can i introduce my network in this frame, and i can adjust this net later

qubvel commented 1 month ago

Hi @wangluoWSG thanks for the issue! Do you mean you want to create a new network similar to Unet (for example), that leverage common features like different backbones/encoders supporting?

wangluoWSG commented 1 month ago

Yes🥳, maybe I can just use my net in this frame and not use these encoders or features. I think this frame is very convenient for conducting experiments. So if it has this function, it will be excellent. Thanks.

没有翅膀却想飞上天空 @.***

qubvel commented 1 month ago

You can use your own net following the example https://github.com/qubvel/segmentation_models.pytorch/blob/master/examples/binary_segmentation_intro.ipynb

It does not necessarily should be Unet or FPN model from the package.

wangluoWSG commented 1 month ago

ok，thanks 

没有翅膀却想飞上天空 @.***

 

------------------ 原始邮件 ------------------ 发件人: "qubvel/segmentation_models.pytorch" @.>; 发送时间: 2024年6月5日(星期三) 下午4:51 @.>; @.**@.>; 主题: Re: [qubvel/segmentation_models.pytorch] the method of adding new networks (Issue #882)

You can use your own net following the example https://github.com/qubvel/segmentation_models.pytorch/blob/master/examples/binary_segmentation_intro.ipynb

It does not necessarily should be Unet or FPN model from the package.

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.Message ID: @.***>