Sepp suggested in our last meeting to use CIFAR 100 (instead of CIFAR 10), which would also strengthen the paper contribution.
However, we have to find a solution to decrease the number of (selected) classes in order to use the current Malevo approach (which is limited to 10 in the conf matrix).
Sepp suggested in our last meeting to use CIFAR 100 (instead of CIFAR 10), which would also strengthen the paper contribution.
However, we have to find a solution to decrease the number of (selected) classes in order to use the current Malevo approach (which is limited to 10 in the conf matrix).