Open denkle opened 1 year ago
Thank you @denkle for another great contribution!
I made some small tweaks to improve the backwards compatibility with torchmetrics
.
Do you want some extra time to figure out why you obtain lower accuracy or is the PR ready to be merged?
Hi, just wanted to ask if you had time to look into this yet?
The code is working properly though the accuracy is lower than expected based in the results reported in the paper.