FOSTER: Feature Boosting and Compression for Class-Incremental Learning

Benchmark Model Rank Results
incremental-learning-on-imagenet-100-50-classes-5-steps-of-10-classesFOSTER#1Average Incremental Accuracy: 80.22
incremental-learning-on-imagenet-100-50-classes-25-steps-of-2-classesFOSTER#2Average Incremental Accuracy: 69.34
incremental-learning-on-imagenet100-10-stepsFOSTER#2Average Incremental Accuracy: 77.75
incremental-learning-on-cifar-100-50-classes-25-steps-of-2-classesFOSTER#3Average Incremental Accuracy: 63.83
incremental-learning-on-imagenet-10-stepsFOSTER#3Average Incremental Accuracy: 68.34
incremental-learning-on-imagenet-100-50-classes-10-steps-of-5-classesFOSTER#3Average Incremental Accuracy: 77.54
incremental-learning-on-cifar-100-50-classes-10-steps-of-5-classesFOSTER#5Average Incremental Accuracy: 67.95
incremental-learning-on-cifar-100-50-classes-5-steps-of-10-classesFOSTER#5Average Incremental Accuracy: 69.46
incremental-learning-on-cifar100-b0-10steps-of-10-classesFOSTER#5Average Incremental Accuracy: 72.9
incremental-learning-on-cifar100b020step-5classesperstepFOSTER#5Average Incremental Accuracy: 70.65