Continual Generalized Category Discovery on ImageNet-100 benchmark leaderboard · SOTA2 Research

Continual Generalized Category Discovery on ImageNet-100

96.21Accuracy (Offline, All)VC-CGCD

Evaluation Results

Method	Links
VC-CGCD2026.07		96.21	91.1	95.52	69	86.31	94.72	65.3	83.33	93.96	65.6	78.6	92.24	61.55	76.72	91.88	61.55
SimGCD2026.07		96.2	79.67	91.68	19.6	70.23	78.83	18.6	61.9	67.43	23.2	56.67	60.92	22.6	52.9	56.4	21.4
FRoST2026.07		96.2	87.5	92.96	60.2	79.63	83.37	57.2	76.78	77	75.2	66.18	68.65	46.4	63.82	66.4	40.6
GM2026.07		96.2	89.53	95.04	62	82.34	86.93	54.8	77.97	79.17	69.6	72.8	74.65	58	71.08	71.76	65
Happy2026.07		96.16	91.03	95.16	70.4	85.46	93.56	65.2	83.07	92.92	66.67	76.96	93.24	56.6	74.98	92.4	57.56
VanillaGCD2026.07		95.96	70.13	72.92	56.2	69.37	73.47	44.8	68.5	70.63	53.6	65.56	67.85	47.2	64.54	67.44	38.4
MetaGCD2026.07		95.96	75.27	78.2	60.6	73.79	75.93	54.9	69.35	72.2	49.4	67.22	70.1	44.2	66.68	69.31	43
KMeans2026.07		85.56	54.9	57.04	44.2	54.73	56.37	44.9	54.67	56.66	40.8	54.63	56.25	41.7	53.92	56.18	33.6