search

msight-tech / research-ms-loss

MS-Loss: Multi-Similarity Loss for Deep Metric Learning
Other
492 stars 76 forks source link

Cars-196 experiments settings #6

Closed visenzeadam closed 5 years ago

visenzeadam commented 5 years ago

I cannot see any experiments setting for Cars-196, can you show the yaml config also?

When I use the CUB-200's yaml config to train the Cars-196(followed the paper's training split rule, 98 classes), but only got best R@1=78.4%(the embedding size is 512), much lower than 84.1% in your paper.

Thanks, Adam

lenscloth commented 5 years ago

@visenzeadam Did you reproduce the result on Cars-196?

geonm commented 5 years ago

@lenscloth

Hi.

You can easily reproduce the results on CARS196.

Use the large batch size more than 450.

lenscloth commented 5 years ago

@geonm Thanks!

wanghao14 commented 5 years ago

@geonm Hi, how could you use such big batch size? Do you use distributed training?