Tiny-imagenet ECE of cross entropy

[SoongE]

Thanks for your great paper and code.

I have some questions about reproducing paper results.

I ran this code without any changes except the dataset path on tiny-imagenet. However, the ResNet50 cross entropy (baseline) ECE is too high than paper.
In my experiment, the ECE of cross entropy is 18.387 without temperature scaling, while the ECE of the paper is 3.73.

I also used your split and the script in README.

I appreciate your attention to this issue.
Thank you for your time and support.

[SoongE]

I have attached the config and log.
dist_train_rank0.log
config.txt