Unable to use ICarL for domain incremental scenario on my dataset in "paths_benchmark"

[2017csz0006]

from avalanche.models.icarl_resnet import *
from avalanche.training.supervised import ICaRL

my_model=IcarlNet(num_classes=2)
my_feature_extractor=my_model.feature_extractor
my_classifier=my_model.classifier

optimizer=Adam(my_model.parameters(), lr=piku_lr)
criterion = BCEWithLogitsLoss()
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")

cl_strategy = ICaRL(
feature_extractor=my_feature_extractor,
classifier=my_classifier,
optimizer=optimizer,
memory_size=200,
buffer_transform=None,
fixed_memory=True,
train_mb_size=64, train_epochs=30, eval_mb_size=64,
evaluator=eval_plugin,
device=device
)

I am getting error as below:

File ~/anaconda3/envs/my_pytorch/lib/python3.10/site-packages/avalanche/training/supervised/icarl.py:159 in before_training_exp
nb_cl = benchmark.n_classes_per_exp[tid]

AttributeError: 'ClassificationScenario' object has no attribute 'n_classes_per_exp'

[Melika-Ayoughi]

Hey @2017csz0006 did you found a way to solve this issue? I'm getting a similar error with the core50 benchmark.