Giter Site home page Giter Site logo

有关网络结构的问题 about rcan HOT 9 CLOSED

yulunzhang avatar yulunzhang commented on August 14, 2024
有关网络结构的问题

from rcan.

Comments (9)

yulunzhang avatar yulunzhang commented on August 14, 2024

您好,
这个卷积层,主要是让其前面能叠加很多个residual block或者residual group,使得这样的残差结构能够训练起来。这其实是借鉴了SRResNet里面的做法(很多residual block叠加之后,以及在long skip connection之前,加一个卷积层)。
不能去除,去除的话,结果就不行了。

from rcan.

splinter21 avatar splinter21 commented on August 14, 2024

您好,
我还有一个有关学习率的疑问,除去CA模块外的部分,和EDSR的block是一致的,初期学习率是1e-4,在加上CA模块后,CA模块的学习率没有另外设置,请问这个是刚好也是1e-4合适吗,还是说CA模块的学习率无需在意?如果我更改了网络需要更改网络的学习率导致不是1e-4,我是否需要在optimizer代码里另外加上找到CA模块让其中的卷积层的学习率控制在1e-4,采用不同层不同学习率的方案?

from rcan.

yulunzhang avatar yulunzhang commented on August 14, 2024

您好,
无需额外设置CA的学习率,保持和其他部分一样的学习率就行。

from rcan.

splinter21 avatar splinter21 commented on August 14, 2024

您好,
我在用您的默认配置,200epoch时decay,到250epoch时loss已经能下降到3.15,我有没有必要跑完所有的1000epoch?因为GPU资源不足,如果可以的话希望直接进行下一个实验。

from rcan.

yulunzhang avatar yulunzhang commented on August 14, 2024

您好,
如果,你模型没有完全确定下来,你可以只跑这么多epoch。
如果,你的模型确定下来,你想跑为论文准备的结果,我建议你跑完1000个epoch。

from rcan.

splinter21 avatar splinter21 commented on August 14, 2024

Hi!
看到你在arxiv投的新文章《Residual Dense Network for Image Restoration 》,上面有关super resolution的指标比CVPR2018的版本高了不少,但是网络图看上去没有变化,想请教一下这个指标的提升是有什么策略吗?
谢谢!

from rcan.

splinter21 avatar splinter21 commented on August 14, 2024

@yulunzhang 以及F-1与F0之间的conv的必要性这个有通过对照实验验证过吗?如果是换成一个block,等于就是最早的block不加入concat而已。conv的效果肯定没有block好,我认为有点多余。

from rcan.

yulunzhang avatar yulunzhang commented on August 14, 2024

Hi!
看到你在arxiv投的新文章《Residual Dense Network for Image Restoration 》,上面有关super resolution的指标比CVPR2018的版本高了不少,但是网络图看上去没有变化,想请教一下这个指标的提升是有什么策略吗?
谢谢!

Hi,
Two differences:
Input size: 32x32 --> 48x48 (same as EDSR)
Training data: DIV2K --> DIV2K + Flickr2K (same as SRMDNF, D-DBPN)

from rcan.

yulunzhang avatar yulunzhang commented on August 14, 2024

@yulunzhang 以及F-1与F0之间的conv的必要性这个有通过对照实验验证过吗?如果是换成一个block,等于就是最早的block不加入concat而已。conv的效果肯定没有block好,我认为有点多余。

Hi,
This conv you mentioned is intended for another usage. In the current version, I didn't show it. Then, it seems to be useless.

from rcan.

Related Issues (20)

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.