Skip to content
Resolution-invariant Person Re-identification
Branch: master
Clone or download
maosnhehe Update
add trained model of market1501
Latest commit 7d85ce0 Jul 11, 2019
Type Name Latest commit message Commit time
Failed to load latest commit information.
torchreid Update Jul 10, 2019 Update Jul 11, 2019 Update Jul 5, 2019
requirements.txt Add files via upload Jul 1, 2019


Resolution-invariant Person Re-identification accepted by IJCAI2019

This code is mainly encouraged by

To accelerate evaluation (10x faster), you can use cython-based evaluation code (developed by luzai). First cd to torchreid/eval_lib, then do make or python build_ext -i. After that, run python to test if the package is successfully installed.


VR dataset, which can be constructed by downsampling the original dataset. can be download in data/ from, whose password is umzg. For example, data/vr_market1501/query/XXX


python -d market1501 --optim adam --lr 0.0003 --max-epoch 60 --stepsize 20 40 --train-batch 32 --test-batch 100 --save-dir log/RIPR_train --gpu-devices 0


python --evaluate -d market1501 --test-batch 100 --save-dir log/RIPR_train --gpu-devices 0 --testmodel log/RIPR_train/best_model.pth.tar


the model file can be downloaded from , whose password is yov0

You can’t perform that action at this time.