RMM: Reinforced Memory Management for Class-Incremental Learning

Open paper
Benchmark Model Rank Results
incremental-learning-on-imagenet-100-50-classes-10-steps-of-5-classesRMM (ResNet-18)#1Average Incremental Accuracy: 78.47
incremental-learning-on-imagenet-100-50-classes-25-steps-of-2-classesRMM (ResNet-18)#1Average Incremental Accuracy: 76.54
incremental-learning-on-imagenet-500-classes-10-steps-of-50-classesRMM (ResNet-18)#1Average Incremental Accuracy: 67.45
incremental-learning-on-imagenet-500-classes-5-steps-of-100-classesRMM (ResNet-18)#1Average Incremental Accuracy: 69.21
incremental-learning-on-imagenet100-10-stepsRMM (ResNet-18)#1Average Incremental Accuracy: 78.47
incremental-learning-on-cifar-100-50-classes-25-steps-of-2-classesRMM (Modified ResNet-32)#2Average Incremental Accuracy: 66.21
incremental-learning-on-imagenet-100-50-classes-5-steps-of-10-classesRMM (ResNet-18)#2Average Incremental Accuracy: 79.52
incremental-learning-on-imagenet-10-stepsRMM (ResNet-18)#4Average Incremental Accuracy: 67.45
incremental-learning-on-cifar-100-50-classes-10-steps-of-5-classesRMM (Modified ResNet-32)#6Average Incremental Accuracy: 67.61
incremental-learning-on-cifar-100-50-classes-5-steps-of-10-classesRMM (Modified ResNet-32)#6Average Incremental Accuracy: 68.86

Rank counts only results with a code link.