GithubHelp home page GithubHelp logo

Comments (9)

yxgeee avatar yxgeee commented on July 17, 2024
  1. 因为在我的实验中, query不做region划分, 只有第一行对应的是query原图与positive的原图及region的相似度

from openibl.

yxgeee avatar yxgeee commented on July 17, 2024
  1. 我不清楚你的query和gallery具体如何划分, 你的query和gallery是否会有类别重叠? 如果你每张query最多只有4个positive, 那么pos_num最大为4, neg_num的话不一定是10, 可以就性能和GPU memory来做具体限制.

from openibl.

yxgeee avatar yxgeee commented on July 17, 2024

另外,如果你在reID上做实验的话,其实我建议在一个reID的codebase上加上这个region的loss. 因为这个code针对的是街景图,无论是图像预处理,还是training scheme都跟reID上最适合的不一样. netvlad也不一定在reID上能收敛的好.

from openibl.

hewumars avatar hewumars commented on July 17, 2024

好的,感谢。我做局部特征的检索,我试验下直接用区域-区域之间是否可以通过这种无监督方式来做。netVLAD刚好也可以用提局部特征。

from openibl.

hewumars avatar hewumars commented on July 17, 2024

还想请教下,sync_gather的两种模式,内存和显存占用有大概统计下极限吗? True的时候11G显存超,False的时候128GB内存超。query+gallery大概5w多张样本。

from openibl.

yxgeee avatar yxgeee commented on July 17, 2024

超内存和超显存的代码位置应该不一样. 为false的时候哪句话超的内存?

from openibl.

hewumars avatar hewumars commented on July 17, 2024

True: dist.all_gather(all_features, features)显存超
False:bc_features.data.copy_(torch.cat(features))内存超

from openibl.

hewumars avatar hewumars commented on July 17, 2024
bc_features = torch.cat(features).cuda(gpu)
for k in range(world_size):
  bc_features.data.copy_(torch.cat(features))
  dist.broadcast(bt_features,k)      

True显存超可能没法解决,False内存超这里有点疑问?
bc_features = torch.cat(features).cuda(gpu)这里定义在gpu上面,为啥copy_的时候是内存在涨?
copy_是不是应该移到循环外面?

from openibl.

yxgeee avatar yxgeee commented on July 17, 2024

这里features太多的时候确实可能超,应该是代码上有缺陷.
有人做了修改(#6) ,可以参考一下

from openibl.

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.