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

multi query vs single query - your results seem reversed #9

Closed jpainam closed 6 years ago

jpainam commented 6 years ago

Hi! Most paper report multi query evaluation greater than single query evaluation. Your evaluation results show the contrary. Didn't reverse the order of the results? And please confirm if the evaluation code you provided is for single query evaluation or muti query

Thanks and great code.

qiaoguan commented 6 years ago

sorry to reply late, you may understand wrong, its true that the multi query evaluation result is mostly better than the single query evaluation, while in my experiment, the single means that we trained the model by single GPU, not the single query evaluation, and so is for the multi. Hope it helps you!

jpainam commented 6 years ago

I got your point. Thank you.