Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

re-ranking #91

Closed
wzjiang opened this issue Dec 26, 2018 · 2 comments
Closed

re-ranking #91

wzjiang opened this issue Dec 26, 2018 · 2 comments
Labels
new_feature New feature (finished)

Comments

@wzjiang
Copy link

wzjiang commented Dec 26, 2018

Thank you for your great work. Could you provide the re-ranking code？
@wzjiang wzjiang closed this as completed Dec 28, 2018
@KaiyangZhou KaiyangZhou reopened this Mar 15, 2019
@KaiyangZhou KaiyangZhou added the planned_feature Planned feature label Mar 15, 2019
@KaiyangZhou KaiyangZhou added new_feature New feature (finished) and removed planned_feature Planned feature labels May 9, 2019
@KaiyangZhou
Copy link
Owner

The reranking code has been added 5839a02.

It's copied from https://github.com/zhunzhong07/person-re-ranking.

Credit to: @zhunzhong07 @michuanhaohao @huanghoujing

How to use:

  1. If you are using your own script, set rerank=True in engine.run() (test_only also needs to be set to True).
  2. If you are using deep-person-reid/scripts/main.py, do --evaluate --rerank. Note that rerank can only be enabled when test_only=True.

@KaiyangZhou
Copy link
Owner

For sanity check, the performance of resnet50 (available in the model zoo) with and without using rerank is reported as follows

Rerank top1 mAP
False 87.9 70.4
True 89.5 84.5

@ghost ghost mentioned this issue Jul 29, 2019
Closed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
new_feature New feature (finished)
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
2 participants
@KaiyangZhou @wzjiang