Class-Incremental Learning on ImageNet (800 base, 40 novel, test) benchmark leaderboard · SOTA2 Research

Class-Incremental Learning on ImageNet (800 base, 40 novel, test)

57.82Accuracy (All Classes)score fusion (ours) best-Accall

Evaluation Results

Method
score fusion (ours) best-Accall2022.04	57.82	57.8	58.23	58.02
pre-trained model2022.04	57.69	58.03	51.2	54.62
score fusion (ours) best-balanced2022.04	55.66	54.81	71.9	63.36
score fusion (ours) best-Accavg2022.04	53.42	52.28	75.13	63.7
learning-based routing w/ Lrt-bal2022.04	52.82	52.12	66.1	59.11
joint learning (oracle)2022.04	47.63	47.32	53.6	50.46
confidence-based routing2022.04	35.43	33.01	81.5	57.25
RM2022.04	21.89	28.78	80.9	49.84
RM2022.04	21.1	18.12	77.7	47.91
iCaRL2022.04	20.23	19.77	28.9	24.34
iCaRL2022.04	18.12	17.04	38.6	27.82
fine-tuning2022.04	3.86	0	77.2	38.6