Can't reproduce the results in the table3

[sherylwang]

Hi, thanks for providing the nice code. I've tried to reproduce the results in table3 with "bash scripts/table_3_rcifar10.sh wrn_28_10". The highest test acc is 87.42% among twice trial, which is much lower than the result in the paper. My tensorflow version is 1.14, ray0.7, cuda10.0, cudnn7. What would be the reason for the low performance?

[arcelien]

Sorry, that scrupt actually reflects the main results in table 2, not the ablation in table 3. I will rename the scripts which were named according to an old version of the paper.

In table 2, a result of 87.42% acc or 12.52% error is better than the reported 12.82% error for rcifar10 and wrn28.

[sherylwang]

Thanks for your prompt reply. I've checked the reproduced result with the reported result in table2, it indeed matched the performance. @arcelien