search

GNAYUOHZ / ReID-MGN

Simple pytorch unofficial implement of paper: Learning Discriminative Features with Multiple Granularities for Person Re-Identification
235 stars 61 forks source link

Results from default settings #3

Closed bityangke closed 5 years ago

bityangke commented 5 years ago

I have trained and evaluated using your default setting, results below can be taken as a reference by others.

start evaluate
[With    Re-Ranking] mAP: 0.9293 rank1: 0.9483 rank3: 0.9676 rank5: 0.9736 rank10: 0.9795
[Without Re-Ranking] mAP: 0.8544 rank1: 0.9376 rank3: 0.9700 rank5: 0.9783 rank10: 0.9863
bityangke commented 5 years ago

Thanks for sharing first. I found the results were about 1 percent lower than the reported results by you. Did you achieve the best results with your default setting?

GNAYUOHZ commented 5 years ago

Hi. Sorry to reply late.I've been busy with exams recently. The only 2 differences are learning rate 1e-4 or 2e-4, and batchid 16 or 8 or 4. When batchid is 16, I found the result will be lower. You can try batchid=4. Thank you.

bityangke commented 5 years ago

Sorry for late response. Got it. Thanks very much! Closed！