GithubHelp home page GithubHelp logo

ufdn-reid's People

Contributors

qianwen96 avatar

Stargazers

 avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar

Watchers

 avatar

ufdn-reid's Issues

A mistake in code and some questions about paper

Hi,
Thanks for sharing the code. It seems that the cls_token would not update through transform block at:

for i , blk in enumerate(self.blocks_token_only):
cls_tokens = blk(global_feat,local_feat_chunk,cls_tokens_list)

Such that, no matter how many layers are set, the results will be the same with using 1 layer.
I also have some questions about code and paper,

  1. I notice that when using model with pd, you use unshared parameters for global and local branch. My questions is, have you ever try to use the shared features. Furthermore, have you ever do the ablation study on only add unshared classifier on these local features extracted by unshared module parameters?
  2. As described in Sec3.2 of your paper, F^i_p is frozen during traing, I did not find this operation in the code. How did you do that?
  3. I notice that you use a learnable parameter gamma to control the learning of attention block, is it the rezero trick for training? Is it critical for network training? @michuanhaohao

Best Regard!

How to maintain the experimental results for TransReID swin-tiny?

Thanks for your insightful work!

I noticed that in your paper, you report results for TransReID with swin-tiny backbone. Could you tell me where to find this implement? Does it contain the jigsaw branch or JPM? I found it's hard to build this module with swin-tiny backbone, given the fact that for local feature branches, it's a token grouping operation in TransReID rather than channel-wise reducing task in your code.

Looking forward to your kind reply. Thanks!

about .txt

图片
你好,我看三个数据集都这个记事本的要求
但是原始数据集里没有
请问您提供的code能跑吗
还是我理解错了

Inquiry about dependencies and code availability for visualizing activation maps

Hi there,
I hope this message finds you well. I recently had the opportunity to read your paper, and I found the section on vehicle feature activation maps particularly intriguing. I was impressed by the insights and visualizations presented in the paper.

I would like to delve deeper into this aspect of your research and explore the generation and interpretation of vehicle feature activation maps. To do so, I am curious to know the specific dependencies or libraries that were utilized in the visualization process. Could you please provide some information regarding the dependencies required to reproduce the vehicle feature activation maps presented in your paper?

Furthermore, I was wondering if the code implementation for generating these activation maps is publicly available. If so, I would greatly appreciate it if you could share the repository link or any instructions to access the code. This would be immensely helpful for me to further understand and apply the techniques discussed in your research.

Thank you very much for your time and for sharing your valuable work with the community. I eagerly await your response.

about CLS Token

你好,我想请教下
CLS Token这块在论文里面没说吧
我怎么没看到
感觉下面这个分支在论文里面没有体现呢
TDH部分有损失;CDC部分也有损失
下面那个加入类别信息的论述在论文中没看到啊

train_viewpoint2.list

Can you provide the train_viewpoint_2.list file of VeRi_WILD and VehicleID, thank you very much!

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.