[WIP] Improve 3D corruptions support

RobustBench / robustbench

RobustBench: a standardized adversarial robustness benchmark [NeurIPS 2021 Benchmarks and Datasets Track]

https://robustbench.github.io

Other

671 stars 99 forks source link

Closed dedeswim closed 1 year ago

dedeswim commented 2 years ago

This PR improves the support for corruptions in the benchmark function. In particular:

[x] Adds torchvision's AlexNet to the models in the model zoo (it is useful to compute mCE)
[x] Adds the possibility to specify a different directory for the clean data and corrupted data
[ ] It adds the computation of mCE (relative to AlexNet) and saves it to the JSON files after running the benchmarking function
[ ] It adds 3D corruptions in the website leaderboard generation script

max-andr commented 1 year ago

I will close this PR as it's included in a newer one https://github.com/RobustBench/robustbench/pull/135.