@@ -54,7 +54,9 @@ CIFAR-100 will be downloaded automatically when running the experiment.
#### ImageNet-Subset
Our code uses exactly the same dataset splits as [PODNet](https://github.com/arthurdouillard/incremental_learning.pytorch).
You need to put the data of [ILSVRC2012](https://www.image-net.org/) under this folder: <https://gitlab.mpi-klsb.mpg.de/yaoyaoliu/rmm/-/tree/main/data/imagenet>
The folders for the training and validation data should be named as `train` and `val`, respectively.
