search

zzpustc / CC-SAM

This is the implementation of our CVPR'23 paper "Class-Conditional Sharpness-Aware Minimization for Deep Long-Tailed Recognition".
16 stars 4 forks source link

Results difference between the body of the paper and the appendix #5

Closed likanchuan closed 1 year ago

likanchuan commented 1 year ago

Hi! Thanks for sharing your excellent work!

I notice there are different numbers between the main table in the body and the ablation study in the appendix under the same setting(is it?) e.g. CIFAR10-LT(beta: 100) 83.92 in Table 1 VS 82.42 in Table 2 in the Appendix. Did I miss something? Looking forward to your reply!

zzpustc commented 1 year ago

Hi,

Apologies for any confusion caused. The discrepancy between the main text and the appendix arises from the request we received to apply strong augmentation to achieve state-of-the-art performance during the rebuttal phase. Consequently, we adhered to these suggestions. However, we also deemed it important to provide ablation studies without the implementation of strong data augmentation, hence their inclusion in the appendix.

Best

likanchuan commented 1 year ago

Thanks for your response! This makes sense to me.