While running both repositories: IBA and IBA-paper-code,
I realized that the estimated mean and std values (pytorch_IBA_per_sample.ipynb) differ from the values provided in your IBA-paper-code repository.
Can you provide some guidance for estimating the same value provided in the IBA-paper-code repository?
Thank you for sharing your excellent project!
While running both repositories: IBA and IBA-paper-code, I realized that the estimated mean and std values (pytorch_IBA_per_sample.ipynb) differ from the values provided in your IBA-paper-code repository.
Can you provide some guidance for estimating the same value provided in the IBA-paper-code repository?