search

qiaoguan / Person-reid-GAN-pytorch

A Pytorch Implementation of "Unlabeled Samples Generated by GAN Improve the Person Re-identification Baseline in vitro"(ICCV17)
158 stars 49 forks source link

Single query evaluation code not available #7

Closed jpainam closed 3 years ago

jpainam commented 6 years ago

Thank for such code. I reproduced the training and the evaluation and confirm the results to be +-1% equal to the reported results. My result for 24000 generated test Rank@1:0.908551 Rank@5:0.969121 Rank@10:0.981591 mAP:0.787033 These are Multi Query evaluation. You reported Single Query Evaluation in your readme. Can you share the code trick to get the Single Query evaluation? Thank

qiaoguan commented 6 years ago

i do not provide single query evaluation, there results are trained by single or multi GPU, and single GPU training is better!