issues
search
dilyabareeva
/
quanda
A toolkit for quantitative evaluation of data attribution methods.
https://quanda.readthedocs.io
MIT License
33
stars
0
forks
source link
Toy Benchmark Base + rename to QuanDA🎉
#60
Closed
dilyabareeva
closed
5 months ago
dilyabareeva
commented
5 months ago
replacing DownstreamTasks with ToyBenchmarks
Refactoring Subclass Detection accordingly
a full usage script for CIFAR-10 and ResNet18
metrics tested on CUDA - making fixes where appropriate
renaming to Quanda
miscellaneous improvements
dilyabareeva
commented
5 months ago
Hi @gumityolcu, please review.